Using Table Automorphisms for Constructing Character Tables in GAP

THOMAS BREUER
Lehrstuhl D für Mathematik
RWTH, 52056 Aachen, Germany

June 27th, 2004

This note has three aims. First it shows how character table automorphisms can be utilized to construct certain character tables from others using the GAP system [GAP04]; the GAP functions used for that are part of the GAP Character Table Library [Bre12]. Second it documents several constructions of character tables which are contained in the GAP Character Table Library. Third it serves as a testfile for the involved GAP functions.

Contents

1  Overview
2  Theoretical Background
    2.1  Character Table Automorphisms
    2.2  Permutation Equivalence of Character Tables
    2.3  Class Fusions
    2.4  Constructing Character Tables of Certain Isoclinic Groups
    2.5  Isoclinic Double Covers of Almost Simple Groups
    2.6  Characters of Normal Subgroups
3  The Constructions
    3.1  Character Tables of Groups of the Structure M.G.A
    3.2  Character Tables of Groups of the Structure G.S3
    3.3  Character Tables of Groups of the Structure G.22
    3.4  Character Tables of Groups of the Structure 22.G (August 2005)
    3.5  p-Modular Tables of Extensions by p-singular Automorphisms
    3.6  Character Tables of Subdirect Products of Index Two (July 2007)
4  Examples for the Type M.G.A
    4.1  Character Tables of Dihedral Groups
    4.2  An M.G.A Type Example with M noncentral in M.G (May 2004)
    4.3  ATLAS Tables of the Type M.G.A
    4.4  More ATLAS Tables of the Type M.G.A
    4.5  The Character Tables of 42.L3(4).23 and 122.L3(4).23
    4.6  The Character Table of (22 ×F4(2)):2 < B (March 2003)
    4.7  The Character Table of 2.(S3 ×Fi22.2) ≤ 2.B (March 2003)
    4.8  The Character Table of (2 ×2.Fi22):2 < Fi24 (November 2008)
    4.9  The Character Table of S3 ×2.U4(3).22 ≤ 2.Fi22 (September 2002)
    4.10  The Character Table of 4.HS.2 ≤ HN.2 (May 2002)
    4.11  The Character Tables of 4.A6.23, 12.A6.23, and 4.L2(25).23
    4.12  Pseudo Character Tables of the Type M.G.A (May 2004)
    4.13  Some Extra-ordinary p-Modular Tables of the Type M.G.A (September 2005)
5  Examples for the Type G.S3
    5.1  Small Examples
    5.2  ATLAS Tables of the Type G.S3
6  Examples for the Type G.22
    6.1  The Character Table of A6.22
    6.2  ATLAS Tables of the Type G.22 - Easy Cases
    6.3  The Character Table of S4(9).22 (September 2011)
    6.4  The Character Tables of Groups of the Type 2.L3(4).22 (June 2010)
    6.5  The Character Tables of Groups of the Type 6.L3(4).22 (October 2011)
    6.6  The Character Tables of Groups of the Type 2.U4(3).22 (February 2012)
    6.7  The Character Tables of Groups of the Type 41.L3(4).22 (October 2011)
    6.8  The Character Tables of Groups of the Type 42.L3(4).22 (October 2011)
    6.9  The Character Table of Aut(L2(81))
    6.10  The Character Table of O8+(3).22111
7  Examples for the Type 22.G
    7.1  The Character Table of 22.Sz(8)
    7.2  ATLAS Tables of the Type 22.G (September 2005)
    7.3  The Character Table of 22.O8+(3) (March 2009)
    7.4  The Character Table of the Schur Cover of L3(4) (September 2005)
8  Examples of Extensions by p-singular Automorphisms
    8.1  Some p-Modular Tables of Groups of the Type M.G.A
    8.2  Some p-Modular Tables of Groups of the Type G.S3
    8.3  2-Modular Tables of Groups of the Type G.22
9  Examples of Subdirect Products of Index Two
    9.1  Certain Dihedral Groups as Subdirect Products of Index Two
    9.2  The Character Table of (D10 ×HN).2 < M (June 2008)

1  Overview

Several types of constructions of character tables of finite groups from known tables of smaller groups are described in Section 3. Selecting suitable character table automorphisms is an important ingredient of these constructions.
Section 2 collects the few representation theoretical facts on which these constructions are based.
The remaining sections show examples of the constructions in GAP. These examples use the GAP Character Table Library, therefore we load this package first.
    gap> LoadPackage( "ctbllib", "1.1.4" );
    true

2  Theoretical Background

2.1  Character Table Automorphisms

Let G be a finite group, Irr(G) be the matrix of ordinary irreducible characters of G, Cl(G) be the set of conjugacy classes of elements in G, gG the G-conjugacy class of g ∈ G, and
powp :
Cl(G)
Cl(G)
gG
(gp)G
the p-th power map, for each prime integer p.
A table automorphism of G is a permutation σ:Cl(G) → Cl(G) with the properties that χ°σ ∈ Irr(G) holds for all χ ∈ Irr(G) and that σ commutes with powp, for all prime integers p that divide the order of G. Note that for prime integers p that are coprime to the order of G, powp commutes with each σ that permutes Irr(G), since powp acts as a field automorphism on the character values.
In GAP, a character table covers the irreducible characters -a matrix M of character values- as well as the power maps of the underlying group -each power map powp being represented as a list powp of positive integers denoting the positions of the image classes. The group of table automorphisms of a character table is represented as a permutation group on the column positions of the table; it can be computed with the function AutomorphismsOfTable.
In the following, we will mainly use that each group automorphism σ of G induces a table automorphism that maps the class of each element in G to the class of its image under σ.

2.2  Permutation Equivalence of Character Tables

Two character tables with matrices M1, M2 of irreducibles and p-th power maps pow1,p, pow2,p are permutation equivalent if permutations ψ and π of row and column positions of the Mi exist such that [ M1 ]i,j = [ M2 ]i ψ, j π holds for all indices i, j, and such that π·pow2,p = pow1,p ·π holds for all primes p that divide the (common) group order. The first condition is equivalent to the existence of a permutation π such that permuting the columns of M1 with π maps the set of rows of M1 to the set of rows of M2.
π is of course determined only up to table automorphisms of the two character tables, that is, two transforming permutations π1, π2 satisfy that π1 ·π2−1 is a table automorphism of the first table, and π1−1 ·π2 is a table automorphism of the second.
Clearly two isomorphic groups have permutation equivalent character tables.
The GAP library function TransformingPermutationsCharacterTables returns a record that contains transforming permutations of rows and columns if the two argument tables are permutation equivalent, and fail otherwise.
In the example sections, the following function for computing representatives from a list of character tables w.r.t. permutation equivalence will be used. More precisely, the input is either a list of character tables or a list of records which have a component table whose value is a character table, and the output is a sublist of the input.
    gap> RepresentativesCharacterTables:= function( list )
    >    local reps, i, found, r;
    > 
    >    reps:= [];
    >    for i in [ 1 .. Length( list ) ] do
    >      if ForAll( reps, r -> ( IsCharacterTable( r ) and
    >             TransformingPermutationsCharacterTables( list[i], r ) = fail )
    >           or ( IsRecord( r ) and TransformingPermutationsCharacterTables(
    >                                    list[i].table, r.table ) = fail ) ) then
    >        Add( reps, list[i] );
    >      fi;
    >    od;
    >    return reps;
    >    end;;

2.3  Class Fusions

For two groups H, G such that H is isomorphic with a subgroup of G, any embedding ι:H → G induces a class function
fusι :
Cl(H)
Cl(G)
hG
(ι(h))G
the class fusion of H in G via ι. Analogously, for a normal subgroup N of G, any epimorphism π:G → G/N induces a class function
fusπ :
Cl(G)
Cl(G/N)
gG
(π(g))G
the class fusion of G onto G/N via π.
When one works only with character tables and not with groups, these class fusions are the objects that describe subgroup and factor group relations between character tables. Technically, class fusions are necessary for restricting, inducing, and inflating characters from one character table to another. If one is faced with the problem to compute the class fusion between the character tables of two groups H and G for which it is known that H can be embedded into G then one can use character-theoretic necessary conditions, concerning that the restriction of all irreducible characters of G to H (via the class fusion) must decompose into the irreducible characters of H, and that the class fusion must commute with the power maps of H and G.
With this character-theoretic approach, one can clearly determine possible class fusions only up to character table automorphisms. Note that one can interpret each character table automorphism of G as a class fusion from the table of G to itself.
If N is a normal subgroup in G then the class fusion of N in G determines the orbits of the conjugation action of G on the classes of N. Often the knowledge of these orbits suffices to identify the subgroup of table automorphisms of N that corresponds to this action of G; for example, this is always the case if N has index 2 in G.
GAP library functions for dealing with class fusions, power maps, and character table automorphisms are described in the chapter "Maps Concerning Character Tables" in the GAP Reference Manual.

2.4  Constructing Character Tables of Certain Isoclinic Groups

As is stated in [CCN+85,p. xxiii], two groups G, H are called isoclinic if they can be embedded into a group K such that K is generated by Z(K) and G, and also by Z(K) and H. In the following, two special cases of isoclinism will be used, where the character tables of the isoclinic groups are closely related.
ctblcons01.png
Starting from the group K containing both G and H, we first note that each irreducible representation of G or H extends to K. More specifically, if ρG is an irreducible representation of G then we can define an extension ρ of K by defining it suitably on Z(K) and then form ρH, the restriction of ρ to H.
In our two cases, we set S = G ∩H, so K = S ∪G \S ∪H \S ∪z S holds for some element z ∈ Z(K) \( G ∪H ) of order four, and G = S ∪g S for some g ∈ G \S, and H = S ∪h S where h = z ·g ∈ H \S. For defining ρH, it suffices to consider ρ(h) = ρ(z) ρ(g), where ρ(z) = ερ(z) ·I is a scalar matrix.
As for the character table heads of G and H, we have sG = sH and z (g ·s)G = (h ·s)H for each s ∈ S, so this defines a bijection of the conjugacy classes of G and H. For a prime integer p, (h ·s)p = (z ·g ·s)p = zp ·(g ·s)p holds for all s ∈ S, so the p-th power maps of G and H are related as follows: Inside S they coincide for any p. If p ≡ 1 mod 4 they coincide also outside S, if p ≡ −1 mod 4 the images differ by exchanging the classes of (h ·s)p and z2 ·(h ·s)p (if these elements lie in different classes), and for p = 2 the images (which lie inside S) differ by exchanging the classes of (h ·s)2 and z2 ·(g ·s)2 (if these elements lie in different classes).
Let ρ be an irreducible representation of K. Then ρG and ρH are related as follows: ρG(s) = ρH(s) and ρ(z) ·ρG(g ·s) = ρH(h ·s) for all s ∈ S. If χG and χH are the characters afforded by ρG and ρH, respectively, then χG(s) = χH(s) and ερ(z) ·χG(g ·s) = χH(h ·s) hold for all s ∈ S. In the case χG(z2) = χ(1) we have ερ(z) = ±1, and both cases actually occur if one considers all irreducible representations of K. In the case χG(z2) = − χ(1) we have ερ(z) = ±i, and again both cases occur. So we obtain the irreducible characters of H from those of G by multiplying the values outside S in all those characters by i that do not have z2 in their kernels.
In GAP, the function CharacterTableIsoclinic can be used for computing the character table of H from that of G, and vice versa. (Note that in the above two cases, also the groups U and H are isoclinic by definition, but CharacterTableIsoclinic does not transfer the character table of U to that of H.)

2.5  Isoclinic Double Covers of Almost Simple Groups

CharacterTableIsoclinic can also be used to switch between the character tables of double covers of groups of the type G.2, where G is a perfect group, see [CCN+85,Section 6.7]. Typical examples are the double covers of symmetric groups.
Note that these double covers may be isomorphic. This happens for 2.S6. More generally, this happens for all semilinear groups ΣL(2,p2), for odd primes p. The smallest examples are ΣL(2,9) = 2.A6.21 and ΣL(2,25) = 2.L2(25).22. This implies that the character table and its isoclinic variant are permutation isomorphic.
    gap> t:= CharacterTable( "2.A6.2_1" );
    CharacterTable( "2.A6.2_1" )
    gap> TransformingPermutationsCharacterTables( t,
    >        CharacterTableIsoclinic( t ) );
    rec( columns := (4,6)(5,7)(11,12)(14,16)(15,17), 
      group := Group([ (16,17), (14,15) ]), rows := (3,5)(4,6)(10,11)(12,15,13,
        14) )
    gap> t:= CharacterTable( "2.L2(25).2_2" );
    CharacterTable( "2.L2(25).2_2" )
    gap> TransformingPermutationsCharacterTables( t,
    >        CharacterTableIsoclinic( t ) );
    rec( columns := (7,9)(8,10)(20,21)(23,24)(25,27)(26,28), 
      group := <permutation group with 4 generators>, 
      rows := (3,5)(4,6)(14,15)(16,17)(19,22,20,21) )

For groups of the type 4.G.2, two different situations can occur. Either the distinguished central cyclic subgroup of order four in 4.G is inverted by the elements in 4.G.2 \4.G, or this subgroup is central in 4.G.2. In the first case, calling CharacterTableIsoclinic with the character table of 4.G.2 yields a character table with the same set of irreducibles, only the 2-power map will in general differ from that of the input table. In the second case, the one argument version of CharacterTableIsoclinic returns a permutation isomorphic table. By supplying additional arguments, there is a chance to construct tables of different groups.
We demonstrate this phenomenon with the various groups of the structure 4.L3(4).2.
    gap> tbls:= [];;
    gap> for m in [ "4_1", "4_2" ] do
    >      for a in [ "2_1", "2_2", "2_3" ] do
    >        Add( tbls, CharacterTable( Concatenation( m, ".L3(4).", a ) ) );
    >      od;
    >    od;
    gap> tbls;
    [ CharacterTable( "4_1.L3(4).2_1" ), CharacterTable( "4_1.L3(4).2_2" ), 
      CharacterTable( "4_1.L3(4).2_3" ), CharacterTable( "4_2.L3(4).2_1" ), 
      CharacterTable( "4_2.L3(4).2_2" ), CharacterTable( "4_2.L3(4).2_3" ) ]
    gap> case1:= Filtered( tbls, t -> Size( ClassPositionsOfCentre( t ) ) = 2 );
    [ CharacterTable( "4_1.L3(4).2_1" ), CharacterTable( "4_1.L3(4).2_2" ), 
      CharacterTable( "4_2.L3(4).2_1" ), CharacterTable( "4_2.L3(4).2_3" ) ]
    gap> case2:= Filtered( tbls, t -> Size( ClassPositionsOfCentre( t ) ) = 4 );
    [ CharacterTable( "4_1.L3(4).2_3" ), CharacterTable( "4_2.L3(4).2_2" ) ]

The centres of the groups 41.L3(4).21, 41.L3(4).22, 42.L3(4).21, and 42.L3(4).23 have order two, that is, these groups belong to the first case. Each of these groups is not permutation equivalent to its isoclinic variant but has the same irreducible characters.
    gap> isos1:= List( case1, CharacterTableIsoclinic );;
    gap> List( [ 1 .. 4 ], i -> Irr( case1[i] ) = Irr( isos1[i] ) );
    [ true, true, true, true ]
    gap> List( [ 1 .. 4 ],
    >      i -> TransformingPermutationsCharacterTables( case1[i], isos1[i] ) );
    [ fail, fail, fail, fail ]

The groups 41.L3(4).23 and 42.L3(4).22 belong to the second case because their centres have order four.
    gap> isos2:= List( case2, CharacterTableIsoclinic );;
    gap> List( [ 1, 2 ],
    >      i -> TransformingPermutationsCharacterTables( case2[i], isos2[i] ) );
    [ rec( columns := (26,27,28,29)(30,31,32,33)(38,39,40,41)(42,43,44,45), 
          group := <permutation group with 5 generators>, 
          rows := (16,17)(18,19)(20,21)(22,23)(28,29)(32,33)(36,37)(40,41) ), 
      rec( columns := (28,29,30,31)(32,33)(34,35,36,37)(38,39,40,41)(42,43,44,
            45)(46,47,48,49), group := <permutation group with 3 generators>, 
          rows := (15,16)(17,18)(20,21)(22,23)(24,25)(26,27)(28,29)(34,35)(38,
            39)(42,43)(46,47) ) ]
    gap> isos3:= List( case2, t -> CharacterTableIsoclinic( t,
    >                                ClassPositionsOfCentre( t ) ) );;
    gap> List( [ 1, 2 ],
    >      i -> TransformingPermutationsCharacterTables( case2[i], isos3[i] ) );
    [ fail, fail ]

2.6  Characters of Normal Subgroups

Let G be a group and N be a normal subgroup of G. We will need the following well-known facts about the relation between the irreducible characters of G and N.
For an irreducible (Brauer) character χ of N and g ∈ G, we define χg by χg(n) = χ(ng) for all n ∈ N, and set IG(χ) = { g ∈ G; χg = χ} (see [Fei82,p. 86]).
If IG(χ) = N then the induced character χG is an irreducible (Brauer) character of G (see [Fei82,Lemma III 2.11]).
If G/N is cyclic and if IG(χ) = G then χ = ψN for an irreducible (Brauer) character ψ of G, and each irreducible (Brauer) character θ with the property χ = θN is of the form θ = ψ·ε, where ε is an irreducible (Brauer) character of G/N (see [Fei82,Theorem III 2.14]).
Clifford's theorem [Fei82,Theorem III 2.12] states that the restriction of an irreducible (Brauer) character of G to N has the form e ∑i=1t ϕi for a positive integer e and irreducible (Brauer) characters ϕi of N, where t is the index of IG1) in G.
Now assume that G is a normal subgroup in a larger group H, that G/N is an abelian chief factor of H and that ψ is an ordinary irreducible character of G such that IH(ψ) = H. Then either t = 1 and e2 is one of 1, |G/N|, or t = |G/N| and e = 1 (see [Isa76,Theorem 6.18]).

3  The Constructions

3.1  Character Tables of Groups of the Structure M.G.A

Let N denote a downward extension of the finite group G by a finite group M, let H denote an automorphic (upward) extension of N by a finite cyclic group A such that M is normal in H, and set F = H / M. We consider the situation that each irreducible character of N that does not contain M in its kernel induces irreducibly to H. Equivalently, the action of A = 〈a 〉 on the characters of N, via χ→ χa, has orbits of length exactly |A| on the set { χ ∈ Irr(N); M ⊄ eq ker(χ) }.
ctblcons02.png
This occurs for example if M is central in N and A acts fixed-point freely on M, we have |M| ≡ 1 mod |A| in this case. If M has prime order then it is sufficient that A does not centralize M.
The ordinary (or p-modular) irreducible characters of H are then given by the ordinary (or p-modular) irreducible characters of F and N, the class fusions from the table of N onto the table of G and from the table of G into that of F, and the permutation π that is induced by the action of A on the conjugacy classes of N.
In general, the action of A on the classes of M is not the right thing to look at, one really must consider the action on the relevant characters of M.G. For example, take H the quaternion group or the dihedral group of order eight, N a cyclic subgroup of index two, and M the centre of H; here A acts trivially on M, but the relevant fact is that the action of A swaps those two irreducible characters of N that take the value −1 on the involution in M -these are the faithful irreducible characters of N.
If the orders of M and A are coprime then also the power maps of H can be computed from the above data. For each prime p that divides the orders of both M and A, the p-th power map is in general not uniquely determined by these input data. In this case, we can compute the (finitely many) candidates for the character table of H that are described by these data. One possible reason for ambiguities is the existence of several isoclinic but nonisomorphic groups that can arise from the input tables (cf. Section 2.4, see Section 4.10 for an example).
With the GAP function PossibleActionsForTypeMGA, one can compute the possible orbit structures induced by G.A on the classes of M.G, and PossibleCharacterTablesForTypeMGA computes the possible ordinary character tables for a given orbit structure. For constructing the p-modular Brauer table of a group H of the structure M.G.A, the GAP function BrauerTableOfTypeMGA takes the ordinary character table of H and the p-modular tables of the subgroup M.G and the factor group G.A as its input. The p-modular table of G is not explicitly needed in the construction, it is implicitly given by the class fusions from M.G into M.G.A and from M.G.A onto G.A; these class fusions must of course be available.
The GAP Character Table Library contains many tables of groups of the structure M.G.A as described above, which are encoded by references to the tables of the groups M.G and G.A, plus the fusion and action information. This reduces the space needed for storing these character tables.
For examples, see Section 4.

3.2  Character Tables of Groups of the Structure G.S3

Let G be a finite group, and H be an upward extension of G such that the factor group H / G is a Frobenius group F = K C with abelian kernel K and cyclic complement C of prime order c. (Typical cases for F are the symmetric group S3 on three points and the alternating group A4 on four points.) Let N and U denote the preimages of K and C under the natural epimorphism from H onto F.
ctblcons03.png
For certain isomorphism types of F, the ordinary (or p-modular) character table of H can be computed from the ordinary (or p-modular) character tables of G, U, and N, the class fusions from the table of G into those of U and N, and the permutation π induced by H on the conjugacy classes of N. This holds for example for F = S3 and in the ordinary case also for F = A4.
Each class of H is either a union of π-orbits or an H-class of U \G; the latter classes are in bijection with the U-classes of U \G, they are just |K| times larger since the |K| conjugates of U in H are fused. The power maps of H are uniquely determined from the power maps of U and N, because each element in F lies in K or in an F-conjugate of C.
Concerning the computation of the ordinary irreducible characters of H, we could induce the irreducible characters of U and N to H, and then take the union of the irreducible characters among those and the irreducible differences of those. (For the case F = S3, this approach has been described in the Appendix of [HL94].)
The GAP function CharacterTableOfTypeGS3 proceeds in a different way, which is suitable also for the construction of p-modular character tables of H.
By the facts listed in Section 2.6, for an irreducible (Brauer) character χ of N, we have IH(χ) equal to either N or H. In the former case, χ induces irreducibly to H. In the latter case, there are extensions ψ(i), 1 ≤ i ≤ |C| (or |C|p′), to H, and we have the following possibilities, depending on the restriction χG.
If χG = e ϕ, for an irreducible character ϕ of G, then IU(ϕ) = U holds, hence the ψ(i)U are |C| (or |C|p′) extensions of χG to U. Moreover, we have either e = 1 or e2 = |K|. In the case e = 1, this determines the values of the ψ(i) on the classes of U outside G. In the case e ≠ 1, we have the problem to combine e extensions of ϕ to a character of U that extends to H.
(One additional piece of information in the case of ordinary character tables is that the norm of this linear combination equals 1 + (|K|−1)/|C|, which determines the ψ(i)U if F = A4 ≅ 22:3 or F = 23:7 holds; in the former case, the sum of each two out of the three different extensions of ϕ extends to U; in the latter case, the sum of all different extensions plus one of the extensions extends. Note that for F = S3, the case e ≠ 1 does not occur.)
The remaining case is that χG is not a multiple of an irreducible character of G. Then χG = ϕ1 + ϕ2 + …+ ϕ|K|, for pairwise different irreducible characters ϕi, 1 ≤ i ≤ |K|, of G with the property ϕiN = χ. The action of U on G fixes at least one of the ϕi, since |K| ≡ 1 mod |C|. Without loss of generality, let IU1) = U, and let [(ϕ1)\tilde](i), 1 ≤ i ≤ |C|, be the extensions of ϕ1 to U. (In fact exactly ϕ1 is fixed by U since otherwise k ∈ K would exist with ϕ1k ≠ ϕ1 and such that also ϕ1k would be invariant in U; but then ϕ1 would be invariant under both C and Ck, which generate F. So each of the |K| constituents is invariant in exactly one of the |K| subgroups of type U above G.)
Then (([(ϕ1)\tilde](i))H)N = ϕ1N = χ, hence the values of ψ(i) on the classes of U \G are given by those of ([(ϕ1)\tilde](i))H. (These are exactly the values of [(ϕ1)\tilde](i). So in both cases, we take the values of χ on N, and on the classes of U \G the values of the extensions of the unique extendible constituent of χG.)
For examples, see Section 5.

3.3  Character Tables of Groups of the Structure G.22

Let G be a finite group, and H be an upward extension of G such that the factor group H / G is a Klein four group. We assume that the ordinary character tables of G and of the three index two subgroups U1, U2, and U3 (of the structures G.21, G.22, and G.23, respectively) of H above G are known, as well as the class fusions of G into these groups. The idea behind the method that is described in this section is that in this situation, there are only few possibilities for the ordinary character table of H.
ctblcons04.png
Namely, the action of H on the classes of G.2i is given by a table automorphism πi of G.2i, and H realizes compatible choices of such automorphisms π1, π2, π3 in the sense that the orbits of all three πi on the classes of G inside the groups G.2i coincide. Furthermore, if G.2i has ni conjugacy classes then an action πi that is a product of fi disjoint transpositions leads to a character table candidate for G.22 that has 2 ni − 3 fi classes, so also the fi must be compatible.
Taking the "inner" classes, i.e., the orbit sums of the classes inside G under the πi, plus the union of the πi-orbits of the classes of G.2i \G gives a possibility for the classes of H. Furthermore, the power maps of the groups G.2i determine the power maps of the candidate table constructed this way.
Concerning the computation of the irreducible characters of H, we consider also the case of p-modular characters tables, where we assume that the ordinary character table of H is already known and the only task is to compute the irreducible p-modular Brauer characters.
Let χ be an irreducible (p-modular Brauer) character of G. By the facts that are listed in Section 2.6, there are three possibilities.
In the latter part of case 3. (except if p = 2), the problem is to combine the values of six irreducible characters of the groups G.2i to four characters of H. This yields essentially two choices, and we try to exclude one possibility by forming scalar products with the 2-nd symmetrizations of the known irreducibles. If several possibilities remain then we get several possible tables.
So we end up with a list of possible character tables of H.
The first step is to specify a list of possible triples (π1, π2, π3), using the table automorphisms of the groups G.2i; this can be done using the GAP function PossibleActionsForTypeGV4. Then the GAP function PossibleCharacterTablesOfTypeGV4 can be used for computing the character table candidates for each given triple of permutations; it may of course happen that some triples of automorphisms are excluded in this second step.
For examples, see Section 6.

3.4  Character Tables of Groups of the Structure 22.G (August 2005)

Let G be a finite group, and H be a central extension of G by a Klein four group Z = 〈z1, z2 〉; set z3 = z1 z2 and Zi = 〈zi 〉, for 1 ≤ i ≤ 3. We assume that the ordinary character tables of the three factor groups 2i.G = H / Zi of H are known, as well as the class fusions from these groups to G. The idea behind the method described in this section is that in this situation, there are only few possibilities for the ordinary character table of H.
ctblcons05.png
Namely, the irreducible (p-modular) characters of H are exactly the inflations of the irreducible (p-modular) characters of the three factor groups H / Zi. (Note that for any noncyclic central subgroup C of H and any χ ∈ Irr(H), we have |ker(χ) ∩C| > 1. To see this, let N = ker(χ). Then clearly |N| > 1, and χ can be regarded as a faithful irreducible character of H/N. If N ∩C would be trivial then N C / N ≅ C would be a noncyclic central subgroup of H/N. This cannot happen by [Isa76,Thm. 2.32 (a)], so the statement can be regarded as an obvious refinement of this theorem.) So all we have to construct is the character table head of H -classes and power maps- and the factor fusions from H to these groups.
For fixed h ∈ H, we consider the question in which H-classes the elements h, h z1, h z2, and h z3 lie. There are three possibilities.
We observe that the question which case actually applies for h ∈ H can be decided from the three factor fusions from H/Zi to G. So we attempt to construct the table head of H and the three factor fusions from H to the groups H/Zi, as follows. Each class gG of G yields either one or two or four preimage classes in H.
In case 1., we get one preimage class in H, and have no choice for the factor fusions.
In case 2., we get two preimage classes, there is exactly one group H/Zi in which gG has two preimage classes -which are in bijection with the two preimage classes of H- and for the other two groups H/Zj, the factor fusions from H map the two classes of H to the unique preimage class of gG. (In the following picture, this is shown for i = 1.)
ctblcons06.png
In case 3., the three factor fusions are in general not uniquely determined: We get four classes, which are defined as two pairs of preimages of the two preimages of gG in H/Z1 and in H/Z2 - so we choose the relevant images in the two factor fusions to H/Z1 and H/Z2, respectively. Note that the class of h in H is the unique class that maps to the class of h Z1 in H/Z1 and to the class of h Z2 in H/Z2, and so on, and we define four classes of H via the four possible combinations of image classes in H/Z1 and H/Z2 (see the picture below).
ctblcons07.png
Due to the fact that in general we do not know which of the two preimage classes of gG in H/Z3 is the class of h Z3, there are in general the following two possibilities for the fusion from H to H/Z3.
ctblcons08.png
This means that we can inflate the irreducible characters of H/Z1 and of H/Z2 to H but that for the inflations of those irreducible characters of H/Z3 to H that are not characters of G, the values on classes where case 3. applies are determined only up to sign.
The GAP function PossibleCharacterTablesOfTypeV4G computes the candidates for the table of H from the tables of the groups H/Zi by setting up the character table head of H using the class fusions from H/Z1 and H/Z2 to G, and then forming the possible class fusions from H to H/Z3.
If case 3. applies for a class gG with g of odd element order then exactly one preimage class in H has odd element order, and we can identify this class in the groups H/Zi, which resolves the ambiguity in this situation. More generally, if g = k2 holds for some k ∈ G then all preimages of kG in H square to the same class of H, so again this class can be identified. In fact PossibleCharacterTablesOfTypeV4G checks whether the p-th power maps of the candidate table for H and the p-th power map of H/Z3 together with the fusion candidate form a commutative diagram.
An additional criterion used by PossibleCharacterTablesOfTypeV4G is given by the property that the product of two characters inflated from H/Z1 and H/Z2, respectively, that are not characters of G is a character of H that contains Z3 in its kernel, so it is checked whether the scalar products of these characters with all characters that are inflated from H/Z3 via the candidate fusion are nonnegative integers.
Once the fusions from H to the groups H/Zi are known, the computation of the irreducible p-modular characters of H from those of the groups H/Zi is straightforward.
The only open question is why this construction is described in this note. That is, how is it related to table automorphisms?
The answer is that in several interesting cases, the three subgroups Z1, Z2, Z3 are conjugate under an order three automorphism σ, say, of H. In this situation, the three factor groups 2i.G = H/Zi are isomorphic, and we can describe the input tables and fusions by the character table of 21.G, the factor fusion from this group to G, and the automorphism [σ] of G that is induced by σ. Assume that σ(Z1) = Z2 holds, and choose h ∈ H. Then σ(h Z1) = σ(h) Z2 is mapped to σ(h) Z = [σ](h Z) under the factor fusion from 22.G to G. Let us start with the character table of 21.G, and fix the class fusion to the character table of G. We may choose the identity map as isomorphism from the table of 21.G to the tables of 22.G and 23.G, which implies that the class of h Z1 is identified with the class of h Z2 and in turn the class fusion from the table of 22.G to that of G can be chosen as the class fusion from the table of 21.G followed by the permutation of classes of G induced by [σ]; analogously, the fusion from the table of 23.G is obtained by applying this permutation twice to the class fusion from the table of 21.G.
For examples, see Section 7.

3.5  p-Modular Tables of Extensions by p-singular Automorphisms

Let G be a finite group, and H be an upward extension of G by an automorphism of prime order p, say. H induces a table automorphism of the p-modular character table of G; let π denote the corresponding permutation of classes of G. The columns of the p-modular character table of H are given by the orbits of π, and the irreducible Brauer characters of H are exactly the orbit sums of π on the irreducible Brauer characters of G.
Note that for computing the p-modular character table of H from that of G, it is sufficient to know the orbits of π and not π itself. Also the ordinary character table of H is not needed, but since GAP stores Brauer character tables relative to their ordinary tables, we are interested mainly in cases where the ordinary character tables of G and H and the p-modular character table of G are known. Assuming that the class fusion between the ordinary tables of G and H is stored on the table of G, the orbits of the action of H on the p-regular classes of G can be read off from it.
The GAP function IBrOfExtensionBySingularAutomorphism can be used to compute the p-modular irreducibles of H.
For examples, see Section 8.

3.6  Character Tables of Subdirect Products of Index Two (July 2007)

Let C2 denote the cyclic group of order two, let G1, G2 be two finite groups, and for i ∈ { 1, 2 }, let ϕi:Gi → C2 be an epimorphism with kernel Hi. Let G be the subdirect product (pullback) of G1 and G2 w.r.t. the epimorphisms ϕi, i.e.,
G = { (g1, g2) ∈ G1 ×G2; ϕ1(g1) = ϕ2(g2) } .
The group G has index two in the direct product G1 ×G2, and G contains H1 ×H2 as a subgroup of index two.
In the following, we describe how the ordinary (or p-modular) character table of G can be computed from the ordinary (or p-modular) character tables of the groups Gi and Hi, and the class fusions from Hi to Gi.
(For the case that one of the groups Gi is a cyclic group of order four, an alternative way to construct the character table of G is described in Section 2.4. For the case that one of the groups Gi acts fixed point freely on the nontrivial irreducible characters of Hi, an alternative construction is described in Section 3.1.)
ctblcons09.png
Each conjugacy class of G is either contained in H1 ×H2 or not. In the former case, let hi ∈ Hi and gi ∈ Gi \Hi; in particular, (g1, g2) ∈ G because both ϕ1(g1) and ϕ2(g2) are not the identity. There are four possibilities.
It remains to deal with the G-classes that are not contained in H1 ×H2. Each such class is in fact a conjugacy class of G1 ×G2. Note that two elements g1, g2 ∈ G1 \H1 are G1-conjugate if and only if they are H1-conjugate. (If g1x = g2 for x ∈ G1 \H1 then g1g1 x = g2 holds, and g1 x ∈ H1.) This implies (g1, g2)G1 ×G2 = (g1, g2)H1 ×H2, and thus this class is equal to (g1, g2)G.
The (ordinary or p-modular) irreducible characters of G are given by the restrictions χG of all those irreducible characters χ of G1 ×G2 whose restriction to H1 ×H2 is irreducible, plus the induced characters ϕG, where ϕ runs over all those irreducible characters of H1 ×H2 that do not occur as restrictions of characters of G1 ×G2.
In other words, no irreducible character of H1 ×H2 has inertia subgroup G inside G1 ×G2. This can be seen as follows. Let ϕ be an irreducible character of H1 ×H2. Then ϕ = ϕ1 ·ϕ2, where ϕ1, ϕ2 are irreducible characters of H1 ×H2 with the properties that H2 ⊆ ker(ϕ1) and H1 ⊆ ker(ϕ2). Sloppy speaking, ϕi is an irreducible character of Hi.
There are four possibilities.
For examples, see Section 9.

4  Examples for the Type M.G.A

4.1  Character Tables of Dihedral Groups

Let n = 2k ·m where k is a nonnegative integer and m is an odd integer, and consider the dihedral group D2n of order 2n. Let N denote the derived subgroup of D2n.
If k = 0 then D2n has the structure M.G.A, with M = N and G the trivial group, and A a cyclic group of order two that inverts each element of N and hence acts fixed-point freely on N. The smallest nontrivial example is of course that of D6 ≅ S3.
    gap> tblMG:= CharacterTable( "Cyclic", 3 );;
    gap> tblG:= CharacterTable( "Cyclic", 1 );;
    gap> tblGA:= CharacterTable( "Cyclic", 2 );;
    gap> StoreFusion( tblMG, [ 1, 1, 1 ], tblG );
    gap> StoreFusion( tblG, [ 1 ], tblGA );
    gap> elms:= Elements( AutomorphismsOfTable( tblMG ) );
    [ (), (2,3) ]
    gap> orbs:= [ [ 1 ], [ 2, 3 ] ];;
    gap> new:= PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA, orbs,
    >              "S3" );
    [ rec( MGfusMGA := [ 1, 2, 2 ], table := CharacterTable( "S3" ) ) ]
    gap> Display( new[1].table );
    S3
    
         2  1  .  1
         3  1  1  .
    
           1a 3a 2a
        2P 1a 3a 1a
        3P 1a 1a 2a
    
    X.1     1  1  1
    X.2     1  1 -1
    X.3     2 -1  .

If k > 0 then D2n has the structure M.G.A, with M = N and G a cyclic group of order two such that M.G is cyclic, and A is a cyclic group of order two that inverts each element of M.G and hence acts fixed-point freely on M.G. The smallest nontrivial example is of course that of D8.
    gap> tblMG:= CharacterTable( "Cyclic", 4 );;
    gap> tblG:= CharacterTable( "Cyclic", 2 );;
    gap> tblGA:= CharacterTable( "2^2" );;           
    gap> OrdersClassRepresentatives( tblMG );
    [ 1, 4, 2, 4 ]
    gap> StoreFusion( tblMG, [ 1, 2, 1, 2 ], tblG ); 
    gap> StoreFusion( tblG, [ 1, 2 ], tblGA );      
    gap> elms:= Elements( AutomorphismsOfTable( tblMG ) );
    [ (), (2,4) ]
    gap> orbs:= Orbits( Group( elms[2] ), [ 1 ..4 ] );;
    gap> new:= PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA, orbs,
    >              "order8" );
    [ rec( MGfusMGA := [ 1, 2, 3, 2 ], table := CharacterTable( "order8" ) ),
      rec( MGfusMGA := [ 1, 2, 3, 2 ], table := CharacterTable( "order8" ) ) ]

Here we get two possible tables, which are the character tables of the dihedral and the quaternion group of order eight, respectively.
    gap> List( new, x -> OrdersClassRepresentatives( x.table ) );
    [ [ 1, 4, 2, 2, 2 ], [ 1, 4, 2, 4, 4 ] ]
    gap> Display( new[1].table );
    order8
    
         2  3  2  3  2  2
    
           1a 4a 2a 2b 2c
        2P 1a 2a 1a 1a 1a
    
    X.1     1  1  1  1  1
    X.2     1  1  1 -1 -1
    X.3     1 -1  1  1 -1
    X.4     1 -1  1 -1  1
    X.5     2  . -2  .  .

For each k > 1 and m = 1, we get two possible tables this way, that of the dihedral group of order 2k+1 and that of the generalized quaternion group of order 2k+1.

4.2  An M.G.A Type Example with M noncentral in M.G (May 2004)

The Sylow 7 normalizer in the symmetric group S12 has the structure 7:6 ×S5, its intersection N with the alternating group A12 is of index two, it has the structure (7:3 ×A5):2.
Let M denote the normal subgroup of order 7 in N, let G denote the normal subgroup of the type 3 ×A5 in F = N/M ≅ 3 ×S5, and A = F/G, the cyclic group of order two. Then N has the structure M.G.A, where A acts fixed-point freely on the irreducible characters of M.G = 7:3 ×A5 that do not contain M in their kernels, hence the character table of N is determined by the character tables of M.G and F, and the action of A on M.G.
Note that in this example, the group M is not central in M.G, unlike in most of our examples.
ctblcons10.png
    gap> tblMG:= CharacterTable( "7:3" ) * CharacterTable( "A5" );;
    gap> nsg:= ClassPositionsOfNormalSubgroups( tblMG );
    [ [ 1 ], [ 1, 6 .. 11 ], [ 1 .. 5 ], [ 1, 6 .. 21 ], [ 1 .. 15 ], [ 1 .. 25 ]
     ]
    gap> List( nsg, x -> Sum( SizesConjugacyClasses( tblMG ){ x } ) );
    [ 1, 7, 60, 21, 420, 1260 ]
    gap> tblG:= tblMG / nsg[2];;
    gap> tblGA:= CharacterTable( "Cyclic", 3 ) * CharacterTable( "A5.2" );;
    gap> GfusGA:= PossibleClassFusions( tblG, tblGA );
    [ [ 1, 2, 3, 4, 4, 8, 9, 10, 11, 11, 15, 16, 17, 18, 18 ],
      [ 1, 2, 3, 4, 4, 15, 16, 17, 18, 18, 8, 9, 10, 11, 11 ] ]
    gap> reps:= RepresentativesFusions( Group(()), GfusGA, tblGA );
    [ [ 1, 2, 3, 4, 4, 8, 9, 10, 11, 11, 15, 16, 17, 18, 18 ] ]
    gap> StoreFusion( tblG, reps[1], tblGA );
    gap> acts:= PossibleActionsForTypeMGA( tblMG, tblG, tblGA );
    [ [ [ 1 ], [ 2 ], [ 3 ], [ 4, 5 ], [ 6, 11 ], [ 7, 12 ], [ 8, 13 ],
          [ 9, 15 ], [ 10, 14 ], [ 16 ], [ 17 ], [ 18 ], [ 19, 20 ], [ 21 ],
          [ 22 ], [ 23 ], [ 24, 25 ] ] ]
    gap> poss:= PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA,
    >               acts[1], "A12N7" );
    [ rec( MGfusMGA := [ 1, 2, 3, 4, 4, 5, 6, 7, 8, 9, 5, 6, 7, 9, 8, 10, 11, 12, 
              13, 13, 14, 15, 16, 17, 17 ], table := CharacterTable( "A12N7" ) ) ]

Let us compare the result table with the table of the Sylow 7 normalizer in A12.
    gap> g:= AlternatingGroup( 12 );;
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[1].table,
    >                CharacterTable( Normalizer( g, SylowSubgroup( g, 7 ) ) ) ) );
    true

Since July 2007, an alternative way to construct the character table of N from other character tables is to exploit its structure as a subdirect product of index two in the group 7:6 ×S5, see Section 3.6.
    gap> tblh1:= CharacterTable( "7:3" );;
    gap> tblg1:= CharacterTable( "7:6" );;
    gap> tblh2:= CharacterTable( "A5" );;
    gap> tblg2:= CharacterTable( "A5.2" );;
    gap> subdir:= CharacterTableOfIndexTwoSubdirectProduct( tblh1, tblg1,
    >                 tblh2, tblg2, "(7:3xA5).2" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[1].table,
    >                subdir.table ) );
    true

For storing the table of N in the GAP Character Table Library, the construction as a subdirect product is more suitable, since the "auxiliary table" of the direct product 7:3 ×A5 need not be stored in the library.

4.3   ATLAS Tables of the Type M.G.A

We show the construction of some character tables of groups of the type M.G.A that are contained in the GAP Character Table Library. Each entry in the following input list contains the names of the library character tables of M.G, G, G.A, and M.G.A.
First we consider the situation where G is a simple group or a central extension of a simple group whose character table is shown in the ATLAS, and M and A are cyclic groups such that M is central in M.G.
In the following cases, the character tables are uniquely determined by the input tables. Note that in each of these cases, |A| and |M| are coprime.
    gap> listMGA:= [
    > [ "3.A6",        "A6",        "A6.2_1",        "3.A6.2_1"       ],
    > [ "3.A6",        "A6",        "A6.2_2",        "3.A6.2_2"       ],
    > [ "6.A6",        "2.A6",      "2.A6.2_1",      "6.A6.2_1"       ],
    > [ "6.A6",        "2.A6",      "2.A6.2_2",      "6.A6.2_2"       ],
    > [ "3.A7",        "A7",        "A7.2",          "3.A7.2"         ],
    > [ "6.A7",        "2.A7",      "2.A7.2",        "6.A7.2"         ],
    > [ "3.L3(4)",     "L3(4)",     "L3(4).2_2",     "3.L3(4).2_2"    ],
    > [ "3.L3(4)",     "L3(4)",     "L3(4).2_3",     "3.L3(4).2_3"    ],
    > [ "6.L3(4)",     "2.L3(4)",   "2.L3(4).2_2",   "6.L3(4).2_2"    ],
    > [ "6.L3(4)",     "2.L3(4)",   "2.L3(4).2_3",   "6.L3(4).2_3"    ],
    > [ "12_1.L3(4)",  "4_1.L3(4)", "4_1.L3(4).2_2", "12_1.L3(4).2_2" ],
    > [ "12_1.L3(4)",  "4_1.L3(4)", "4_1.L3(4).2_3", "12_1.L3(4).2_3" ],
    > [ "12_2.L3(4)",  "4_2.L3(4)", "4_2.L3(4).2_2", "12_2.L3(4).2_2" ],
    > [ "12_2.L3(4)",  "4_2.L3(4)", "4_2.L3(4).2_3", "12_2.L3(4).2_3" ],
    > [ "3.U3(5)",     "U3(5)",     "U3(5).2",       "3.U3(5).2"      ],
    > [ "3.M22",       "M22",       "M22.2",         "3.M22.2"        ],
    > [ "6.M22",       "2.M22",     "2.M22.2",       "6.M22.2"        ],
    > [ "12.M22",      "4.M22",     "4.M22.2",       "12.M22.2"       ],
    > [ "3.L3(7)",     "L3(7)",     "L3(7).2",       "3.L3(7).2"      ],
    > [ "3_1.U4(3)",   "U4(3)",     "U4(3).2_1",     "3_1.U4(3).2_1"  ],
    > [ "3_1.U4(3)",   "U4(3)",     "U4(3).2_2'",    "3_1.U4(3).2_2'" ],
    > [ "3_2.U4(3)",   "U4(3)",     "U4(3).2_1",     "3_2.U4(3).2_1"  ],
    > [ "3_2.U4(3)",   "U4(3)",     "U4(3).2_3'",    "3_2.U4(3).2_3'" ],
    > [ "6_1.U4(3)",   "2.U4(3)",   "2.U4(3).2_1",   "6_1.U4(3).2_1"  ],
    > [ "6_1.U4(3)",   "2.U4(3)",   "2.U4(3).2_2'",  "6_1.U4(3).2_2'" ],
    > [ "6_2.U4(3)",   "2.U4(3)",   "2.U4(3).2_1",   "6_2.U4(3).2_1"  ],
    > [ "6_2.U4(3)",   "2.U4(3)",   "2.U4(3).2_3'",  "6_2.U4(3).2_3'" ],
    > [ "12_1.U4(3)",  "4.U4(3)",   "4.U4(3).2_1",   "12_1.U4(3).2_1" ],
    > [ "12_2.U4(3)",  "4.U4(3)",   "4.U4(3).2_1",   "12_2.U4(3).2_1" ],
    > [ "3.G2(3)",     "G2(3)",     "G2(3).2",       "3.G2(3).2"      ],
    > [ "3.U3(8)",     "U3(8)",     "U3(8).2",       "3.U3(8).2"      ],
    > [ "3.U3(8).3_1", "U3(8).3_1", "U3(8).6",       "3.U3(8).6"      ],
    > [ "3.J3",        "J3",        "J3.2",          "3.J3.2"         ],
    > [ "3.U3(11)",    "U3(11)",    "U3(11).2",      "3.U3(11).2"     ],
    > [ "3.McL",       "McL",       "McL.2",         "3.McL.2"        ],
    > [ "3.O7(3)",     "O7(3)",     "O7(3).2",       "3.O7(3).2"      ],
    > [ "6.O7(3)",     "2.O7(3)",   "2.O7(3).2",     "6.O7(3).2"      ],
    > [ "3.U6(2)",     "U6(2)",     "U6(2).2",       "3.U6(2).2"      ],
    > [ "6.U6(2)",     "2.U6(2)",   "2.U6(2).2",     "6.U6(2).2"      ],
    > [ "3.Suz",       "Suz",       "Suz.2",         "3.Suz.2"        ],
    > [ "6.Suz",       "2.Suz",     "2.Suz.2",       "6.Suz.2"        ],
    > [ "3.ON",        "ON",        "ON.2",          "3.ON.2"         ],
    > [ "3.Fi22",      "Fi22",      "Fi22.2",        "3.Fi22.2"       ],
    > [ "6.Fi22",      "2.Fi22",    "2.Fi22.2",      "6.Fi22.2"       ],
    > [ "3.2E6(2)",    "2E6(2)",    "2E6(2).2",      "3.2E6(2).2"     ],
    > [ "6.2E6(2)",    "2.2E6(2)",  "2.2E6(2).2",    "6.2E6(2).2"     ],
    > [ "3.F3+",       "F3+",       "F3+.2",         "3.F3+.2"        ],
    > ];;

Note that the groups of the types 121.L3(4).21 and 122.L3(4).21 have central subgroups of order six, so we cannot choose G equal to 41.L3(4) and 42.L3(4), respectively, in these cases. See Section 4.4 for the construction of these tables.
Also in the following cases, |A| and |M| are coprime, we have |M| = 3 and |A| = 2. The group M.G has a central subgroup of the type 22 ×3, and A acts on this group by inverting the elements in the subgroup of order 3 and by swapping two involutions in the Klein four group.
    gap> Append( listMGA, [
    > [ "(2^2x3).L3(4)",  "2^2.L3(4)",   "2^2.L3(4).2_2", "(2^2x3).L3(4).2_2" ],
    > [ "(2^2x3).L3(4)",  "2^2.L3(4)",   "2^2.L3(4).2_3", "(2^2x3).L3(4).2_3" ],
    > [ "(2^2x3).U6(2)",  "2^2.U6(2)",   "2^2.U6(2).2",   "(2^2x3).U6(2).2"   ],
    > [ "(2^2x3).2E6(2)", "2^2.2E6(2)",  "2^2.2E6(2).2",  "(2^2x3).2E6(2).2"  ],
    > ] );

Additionally, there are a few cases where A has order two, and G.A has a factor group of the type 22, and a few cases where M has the type 22 and A is of order three and acts transitively on the involutions in M.
    gap> Append( listMGA, [
    > [ "3.A6.2_3",       "A6.2_3",    "A6.2^2",      "3.A6.2^2"          ],
    > [ "3.L3(4).2_1",    "L3(4).2_1", "L3(4).2^2",   "3.L3(4).2^2"       ],
    > [ "3_1.U4(3).2_2",  "U4(3).2_2", "U4(3).(2^2)_{122}",
    >                                             "3_1.U4(3).(2^2)_{122}" ],
    > [ "3_2.U4(3).2_3",  "U4(3).2_3", "U4(3).(2^2)_{133}",
    >                                             "3_2.U4(3).(2^2)_{133}" ],
    > [ "3^2.U4(3).2_3'", "3_2.U4(3).2_3'", "3_2.U4(3).(2^2)_{133}",
    >                                             "3^2.U4(3).(2^2)_{133}" ],
    > [ "2^2.L3(4)",      "L3(4)",     "L3(4).3",     "2^2.L3(4).3"       ],
    > [ "(2^2x3).L3(4)",  "3.L3(4)",   "3.L3(4).3",   "(2^2x3).L3(4).3"   ],
    > [ "2^2.L3(4).2_1",  "L3(4).2_1", "L3(4).6",     "2^2.L3(4).6"       ],
    > [ "2^2.Sz(8)",      "Sz(8)",     "Sz(8).3",     "2^2.Sz(8).3"       ],
    > [ "2^2.U6(2)",      "U6(2)",     "U6(2).3",     "2^2.U6(2).3"       ],
    > [ "(2^2x3).U6(2)",  "3.U6(2)",   "3.U6(2).3",   "(2^2x3).U6(2).3"   ],
    > [ "2^2.O8+(2)",     "O8+(2)",    "O8+(2).3",    "2^2.O8+(2).3"      ],
    > [ "2^2.O8+(3)",     "O8+(3)",    "O8+(3).3",    "2^2.O8+(3).3"      ],
    > [ "2^2.2E6(2)",     "2E6(2)",    "2E6(2).3",    "2^2.2E6(2).3"      ],
    > ] );

The constructions of the character tables of groups of the types 42.L3(4).23 and 122.L3(4).23 is described in Section 4.5, in these cases the GAP functions return several possible tables.
The construction of the various character table of groups of the types 41.L3(4).22 and 42.L3(4).22 are described in Section 6.7.
The following function takes the ordinary character tables of the groups M.G, G, and G.A, a string to be used as the Identifier value of the character table of M.G.A, and the character table of M.G.A that is contained in the GAP Character Table Library; the function first computes the possible actions of G.A on the classes of M.G, using the function PossibleActionsForTypeMGA, then computes the union of possible character tables for these actions, and then representatives up to permutation equivalence; if there is only one solution then the result table is compared with the library table.
    gap> ConstructOrdinaryMGATable:= function( tblMG, tblG, tblGA, name, lib )
    >      local acts, poss, trans;
    > 
    >      acts:= PossibleActionsForTypeMGA( tblMG, tblG, tblGA );
    >      poss:= Concatenation( List( acts, pi ->
    >                 PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA, pi,
    >                     name ) ) );
    >      poss:= RepresentativesCharacterTables( poss );
    >      if Length( poss ) = 1 then
    >        # Compare the computed table with the library table.
    >        if not IsCharacterTable( lib ) then
    >          List( poss, x -> AutomorphismsOfTable( x.table ) );
    >          Print( "#I  no library table for ", name, "\n" );
    >        else
    >          trans:= TransformingPermutationsCharacterTables( poss[1].table,
    >                      lib );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " differ\n" );
    >          fi;
    >          # Compare the computed fusion with the stored one.
    >          if OnTuples( poss[1].MGfusMGA, trans.columns )
    >                 <> GetFusionMap( tblMG, lib ) then
    >            Print( "#E  computed and stored fusion for ", name,
    >                   " differ\n" );
    >          fi;
    >        fi;
    >      elif Length( poss ) = 0 then
    >        Print( "#E  no solution for ", name, "\n" );
    >      else
    >        Print( "#E  ", Length( poss ), " possibilities for ", name, "\n" );
    >      fi;
    >      return poss;
    >    end;;

The following function takes the ordinary character tables of the groups M.G, G.A, and M.G.A, and tries to construct the p-modular character tables of M.G.A from the p-modular character tables of the first two of these tables, for all prime divisors p of the order of M.G.A. Note that the tables of G are not needed in the construction, only the class fusions from M.G to M.G.A and from M.G.A to G.A must be stored.
    gap> ConstructModularMGATables:= function( tblMG, tblGA, ordtblMGA )
    >    local name, poss, p, modtblMG, modtblGA, modtblMGA, modlib, trans;
    > 
    >    name:= Identifier( ordtblMGA );
    >    poss:= [];
    >    for p in Set( Factors( Size( ordtblMGA ) ) ) do
    >      modtblMG := tblMG mod p;
    >      modtblGA := tblGA mod p;
    >      if ForAll( [ modtblMG, modtblGA ], IsCharacterTable ) then
    >        modtblMGA:= BrauerTableOfTypeMGA( modtblMG, modtblGA, ordtblMGA );
    >        Add( poss, modtblMGA );
    >        modlib:= ordtblMGA mod p;
    >        if IsCharacterTable( modlib ) then
    >          trans:= TransformingPermutationsCharacterTables( modtblMGA.table,
    >                      modlib );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " mod ", p, " differ\n" );
    >          fi;
    >        else
    >          AutomorphismsOfTable( modtblMGA.table );
    >          Print( "#I  no library table for ", name, " mod ", p, "\n" );
    >        fi;
    >      else
    >        Print( "#I  not all input tables for ", name, " mod ", p,
    >               " available\n" );
    >      fi;
    >    od;
    > 
    >    return poss;
    >    end;;

Now we run the constructions for the cases in the list. Note that in order to avoid conflicts of the class fusions that arise in the construction with the class fusions that are already stored on the library tables, we choose identifiers for the result tables that are different from the identifiers of the library tables.
    gap> for  input in listMGA do
    >      tblMG := CharacterTable( input[1] );
    >      tblG  := CharacterTable( input[2] );
    >      tblGA := CharacterTable( input[3] );
    >      name  := Concatenation( "new", input[4] );
    >      lib   := CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );
    >      if 1 <> Length( poss ) then
    >        Print( "#I  ", Length( poss ), " possibilities for ", name, "\n" );
    >      elif lib = fail then
    >        Print( "#I  no library table for ", input[4], "\n" );
    >      else
    >        ConstructModularMGATables( tblMG, tblGA, lib );
    >      fi;
    >    od;
    #I  not all input tables for 6.Suz.2 mod 13 available
    #I  not all input tables for 3.ON.2 mod 3 available
    #I  not all input tables for 3.2E6(2).2 mod 2 available
    #I  not all input tables for 3.2E6(2).2 mod 3 available
    #I  not all input tables for 3.2E6(2).2 mod 5 available
    #I  not all input tables for 3.2E6(2).2 mod 7 available
    #I  not all input tables for 3.2E6(2).2 mod 11 available
    #I  not all input tables for 3.2E6(2).2 mod 13 available
    #I  not all input tables for 3.2E6(2).2 mod 17 available
    #I  not all input tables for 3.2E6(2).2 mod 19 available
    #I  not all input tables for 6.2E6(2).2 mod 2 available
    #I  not all input tables for 6.2E6(2).2 mod 3 available
    #I  not all input tables for 6.2E6(2).2 mod 5 available
    #I  not all input tables for 6.2E6(2).2 mod 7 available
    #I  not all input tables for 6.2E6(2).2 mod 11 available
    #I  not all input tables for 6.2E6(2).2 mod 13 available
    #I  not all input tables for 6.2E6(2).2 mod 17 available
    #I  not all input tables for 6.2E6(2).2 mod 19 available
    #I  not all input tables for 3.F3+.2 mod 2 available
    #I  not all input tables for 3.F3+.2 mod 3 available
    #I  not all input tables for 3.F3+.2 mod 5 available
    #I  not all input tables for 3.F3+.2 mod 7 available
    #I  not all input tables for 3.F3+.2 mod 13 available
    #I  not all input tables for 3.F3+.2 mod 17 available
    #I  not all input tables for 3.F3+.2 mod 29 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 2 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 3 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 5 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 7 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 11 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 13 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 17 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 19 available
    #I  not all input tables for 3^2.U4(3).(2^2)_{133} mod 2 available
    #I  not all input tables for 3^2.U4(3).(2^2)_{133} mod 5 available
    #I  not all input tables for 3^2.U4(3).(2^2)_{133} mod 7 available
    #I  not all input tables for 2^2.O8+(3).3 mod 3 available
    #I  not all input tables for 2^2.O8+(3).3 mod 5 available
    #I  not all input tables for 2^2.O8+(3).3 mod 7 available
    #I  not all input tables for 2^2.O8+(3).3 mod 13 available
    #I  not all input tables for 2^2.2E6(2).3 mod 2 available
    #I  not all input tables for 2^2.2E6(2).3 mod 3 available
    #I  not all input tables for 2^2.2E6(2).3 mod 5 available
    #I  not all input tables for 2^2.2E6(2).3 mod 7 available
    #I  not all input tables for 2^2.2E6(2).3 mod 11 available
    #I  not all input tables for 2^2.2E6(2).3 mod 13 available
    #I  not all input tables for 2^2.2E6(2).3 mod 17 available
    #I  not all input tables for 2^2.2E6(2).3 mod 19 available

We do not get any unexpected output, so the character tables in question are determined by the inputs.
Alternative constructions of the character tables of 3.A6.22, 3.L3(4).22, and 32.U4(3).(22)133 can be found in Section 6.2.

4.4  More ATLAS Tables of the Type M.G.A

In the following situations, we have |A| = 2, and |M| is a multiple of 2. The result turns out to be unique up to isoclinism, see Section 3.1.
First, there are some cases where the centre of M.G is a cyclic group of order four, and |M| = 2 holds.
    gap> listMGA2:= [
    > [ "4_1.L3(4)",  "2.L3(4)",   "2.L3(4).2_1",   "4_1.L3(4).2_1"  ],
    > [ "4_1.L3(4)",  "2.L3(4)",   "2.L3(4).2_2",   "4_1.L3(4).2_2"  ],
    > [ "4_2.L3(4)",  "2.L3(4)",   "2.L3(4).2_1",   "4_2.L3(4).2_1"  ],
    > [ "4.M22",      "2.M22",     "2.M22.2",       "4.M22.2"        ],
    > [ "4.U4(3)",    "2.U4(3)",   "2.U4(3).2_2",   "4.U4(3).2_2"    ],
    > [ "4.U4(3)",    "2.U4(3)",   "2.U4(3).2_3",   "4.U4(3).2_3"    ],
    > ];;

Note that the groups 41.L3(4).23 and 42.L3(4).22 and their isoclinic variants have centres of order four, so they do not appear here. The construction of the character table of 42.L3(4).23 is more involved, it is described in Section 4.5.
Also in the following cases, we have |M| = 2, but the situation is different because M.G has a central subgroup of the type 22 containing a unique subgroup of order 2 that is central in M.G.A.
    gap> Append( listMGA2, [
    > [ "2^2.L3(4)",     "2.L3(4)",     "2.L3(4).2_2",         "2^2.L3(4).2_2" ],
    > [ "2^2.L3(4)",     "2.L3(4)",     "2.L3(4).2_3",         "2^2.L3(4).2_3" ],
    > [ "2^2.L3(4).2_1", "2.L3(4).2_1", "2.L3(4).(2^2)_{123}", "2^2.L3(4).2^2" ],
    > [ "2^2.O8+(2)",    "2.O8+(2)",    "2.O8+(2).2",          "2^2.O8+(2).2"  ],
    > [ "2^2.U6(2)",     "2.U6(2)",     "2.U6(2).2",           "2^2.U6(2).2"   ],
    > [ "2^2.2E6(2)",    "2.2E6(2)",    "2.2E6(2).2",          "2^2.2E6(2).2"  ],
    > ] );

Next there are two constructions for G = 6.L3(4), with |M| = 12 and |A| = 2. Note that the groups 121.L3(4).21 and 122.L3(4).21 have central subgroups of the order six, so we cannot use the factor groups 41.L3(4).21 and 42.L3(4).21, respectively, for the constructions.
    gap> Append( listMGA2, [
    > [ "12_1.L3(4)", "6.L3(4)", "6.L3(4).2_1", "12_1.L3(4).2_1" ],
    > [ "12_2.L3(4)", "6.L3(4)", "6.L3(4).2_1", "12_2.L3(4).2_1" ],
    > ] );

Next there are alternative constructions for tables which have been constructed in Section 4.3. There we had viewed the groups of the structure 12.S.2, for a simple group S, as 3.G.2 with G = 4.S. Here we view these groups as 2.G.2 with G = 6.S, which means that we do not prescribe the 4.S.2 type factor group. So it is not surprising that we get more than one solution, and that the computation of the 2-power map of 12.S.2 is more involved. Note that the construction of the character table of 122.L3(4).23 is more involved, it is described in Section 4.5.
    gap> Append( listMGA2, [
    > [ "12.M22",     "6.M22",     "6.M22.2",       "12.M22.2"       ],
    > [ "12_1.L3(4)", "6.L3(4)",   "6.L3(4).2_2",   "12_1.L3(4).2_2" ],
    > [ "12_1.U4(3)", "6_1.U4(3)", "6_1.U4(3).2_2", "12_1.U4(3).2_2" ],
    > [ "12_2.U4(3)", "6_2.U4(3)", "6_2.U4(3).2_3", "12_2.U4(3).2_3" ],
    > ] );

Finally, there are alternative constructions for the cases where the group M.G has a central subgroup of the type 22 ×3, and A acts on this group by inverting the elements in the subgroup of order 3 and by swapping two involutions in the Klein four group.
    gap> Append( listMGA2, [
    > [ "(2^2x3).L3(4)",  "6.L3(4)",   "6.L3(4).2_2", "(2^2x3).L3(4).2_2" ],
    > [ "(2^2x3).L3(4)",  "6.L3(4)",   "6.L3(4).2_3", "(2^2x3).L3(4).2_3" ],
    > [ "(2^2x3).U6(2)",  "6.U6(2)",   "6.U6(2).2",   "(2^2x3).U6(2).2"   ],
    > [ "(2^2x3).2E6(2)", "6.2E6(2)",  "6.2E6(2).2",  "(2^2x3).2E6(2).2"  ],
    > ] );

Now we run the constructions for the cases in the list.
    gap> for  input in listMGA2 do
    >      tblMG := CharacterTable( input[1] );
    >      tblG  := CharacterTable( input[2] );
    >      tblGA := CharacterTable( input[3] );
    >      name  := Concatenation( "new", input[4] );
    >      lib   := CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );
    >      if Length( poss ) = 2 then
    >        iso:= CharacterTableIsoclinic( poss[1].table );
    >        if IsRecord( TransformingPermutationsCharacterTables( poss[2].table,
    >                         iso ) ) then
    >          Unbind( poss[2] );
    >        fi;
    >      elif Length( poss ) = 1 then
    >        Print( "#I  unique up to permutation equivalence: ", name, "\n" );
    >      fi;
    >      if 1 <> Length( poss ) then
    >        Print( "#I  ", Length( poss ), " possibilities for ", name, "\n" );
    >      elif lib = fail then
    >        Print( "#I  no library table for ", input[4], "\n" );
    >      else
    >        ConstructModularMGATables( tblMG, tblGA, lib );
    >      fi;
    >    od;
    #E  2 possibilities for new4_1.L3(4).2_1
    #E  2 possibilities for new4_1.L3(4).2_2
    #E  2 possibilities for new4_2.L3(4).2_1
    #E  2 possibilities for new4.M22.2
    #E  2 possibilities for new4.U4(3).2_2
    #E  2 possibilities for new4.U4(3).2_3
    #I  unique up to permutation equivalence: new2^2.L3(4).2_2
    #I  unique up to permutation equivalence: new2^2.L3(4).2_3
    #I  unique up to permutation equivalence: new2^2.L3(4).2^2
    #I  unique up to permutation equivalence: new2^2.O8+(2).2
    #I  unique up to permutation equivalence: new2^2.U6(2).2
    #I  unique up to permutation equivalence: new2^2.2E6(2).2
    #I  not all input tables for 2^2.2E6(2).2 mod 2 available
    #I  not all input tables for 2^2.2E6(2).2 mod 3 available
    #I  not all input tables for 2^2.2E6(2).2 mod 5 available
    #I  not all input tables for 2^2.2E6(2).2 mod 7 available
    #E  2 possibilities for new12_1.L3(4).2_1
    #E  2 possibilities for new12_2.L3(4).2_1
    #E  2 possibilities for new12.M22.2
    #E  2 possibilities for new12_1.L3(4).2_2
    #E  2 possibilities for new12_1.U4(3).2_2
    #E  2 possibilities for new12_2.U4(3).2_3
    #I  unique up to permutation equivalence: new(2^2x3).L3(4).2_2
    #I  unique up to permutation equivalence: new(2^2x3).L3(4).2_3
    #I  unique up to permutation equivalence: new(2^2x3).U6(2).2
    #I  unique up to permutation equivalence: new(2^2x3).2E6(2).2
    #I  not all input tables for (2^2x3).2E6(2).2 mod 2 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 3 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 5 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 7 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 11 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 13 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 17 available
    #I  not all input tables for (2^2x3).2E6(2).2 mod 19 available

Again, we do not get any unexpected output, so the character tables in question are determined up to isoclinism by the inputs.

4.5  The Character Tables of 42.L3(4).23 and 122.L3(4).23

In the construction of the character table of M.G.A = 42.L3(4).23 from the tables of M.G = 42.L3(4) and G.A = 2.L3(4).23, the action of A on the classes of M.G is uniquely determined, but we get four possible character tables.
    gap> tblMG := CharacterTable( "4_2.L3(4)" );;
    gap> tblG  := CharacterTable( "2.L3(4)" );;
    gap> tblGA := CharacterTable( "2.L3(4).2_3" );;
    gap> name  := "new4_2.L3(4).2_3";;
    gap> lib   := CharacterTable( "4_2.L3(4).2_3" );;
    gap> poss  := ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );
    #E  4 possibilities for new4_2.L3(4).2_3
    [ rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 7, 9, 10, 11, 10, 12, 12, 13, 
              14, 15, 14, 16, 17, 18, 17, 19, 20, 21, 22, 19, 22, 21, 20 ], 
          table := CharacterTable( "new4_2.L3(4).2_3" ) ), 
      rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 7, 9, 10, 11, 10, 12, 12, 13, 
              14, 15, 14, 16, 17, 18, 17, 19, 20, 21, 22, 19, 22, 21, 20 ], 
          table := CharacterTable( "new4_2.L3(4).2_3" ) ), 
      rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 7, 9, 10, 11, 10, 12, 12, 13, 
              14, 15, 14, 16, 17, 18, 17, 19, 20, 21, 22, 19, 22, 21, 20 ], 
          table := CharacterTable( "new4_2.L3(4).2_3" ) ), 
      rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 7, 9, 10, 11, 10, 12, 12, 13, 
              14, 15, 14, 16, 17, 18, 17, 19, 20, 21, 22, 19, 22, 21, 20 ], 
          table := CharacterTable( "new4_2.L3(4).2_3" ) ) ]

The centre of 42.L3(4) is inverted by the action of the outer automorphism, so the existence of two possible tables can be expected because two isoclinic groups of the type 42.L3(4).23 exist, see Section 2.5.
Indeed the result consists of two pairs of isoclinic tables, so we have to decide which pair of tables belongs to the groups of the type 42.L3(4).23.
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[1].table,
    >                  CharacterTableIsoclinic( poss[4].table ) ) );
    true
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[2].table,
    >                  CharacterTableIsoclinic( poss[3].table ) ) );
    true

The possible tables differ only w.r.t. the 2-power map and perhaps the element orders. The ATLAS prints the table of the split extension of M.G, this table is one of the first two possibilities.
    gap> List( poss, x -> PowerMap( x.table, 2 ) );
    [ [ 1, 3, 1, 1, 3, 6, 8, 6, 4, 4, 4, 5, 16, 18, 16, 13, 15, 13, 19, 21, 19, 
          21, 1, 1, 6, 6, 9, 9, 11, 11, 16, 16, 13, 13 ], 
      [ 1, 3, 1, 1, 3, 6, 8, 6, 4, 4, 4, 5, 16, 18, 16, 13, 15, 13, 19, 21, 19, 
          21, 1, 1, 6, 6, 11, 11, 9, 9, 16, 16, 13, 13 ], 
      [ 1, 3, 1, 1, 3, 6, 8, 6, 4, 4, 4, 5, 16, 18, 16, 13, 15, 13, 19, 21, 19, 
          21, 3, 3, 8, 8, 9, 9, 11, 11, 18, 18, 15, 15 ], 
      [ 1, 3, 1, 1, 3, 6, 8, 6, 4, 4, 4, 5, 16, 18, 16, 13, 15, 13, 19, 21, 19, 
          21, 3, 3, 8, 8, 11, 11, 9, 9, 18, 18, 15, 15 ] ]

The 2-power map is not determined by the irreducible characters (and by the 2-power map of the factor group 2.L3(4).23). We determine this map using the embedding of 42.L3(4).23 into 4.U4(3).23. Note that L3(4).23 is a maximal subgroup of U4(3).23 (see [CCN+85,p. 52]), and that the subgroup L3(4) of U4(3) lifts to 42.L3(4) in 4.U4(3) because no embedding of L3(4), 2.L3(4), or 41.L3(4) into 4.U4(3) is possible.
    gap> PossiblePowerMaps( poss[1].table, 2 );
    [ [ 1, 3, 1, 1, 3, 6, 8, 6, 4, 4, 4, 5, 16, 18, 16, 13, 15, 13, 19, 21, 19, 
          21, 1, 1, 6, 6, 11, 11, 9, 9, 16, 16, 13, 13 ], 
      [ 1, 3, 1, 1, 3, 6, 8, 6, 4, 4, 4, 5, 16, 18, 16, 13, 15, 13, 19, 21, 19, 
          21, 1, 1, 6, 6, 9, 9, 11, 11, 16, 16, 13, 13 ] ]
    gap> t:= CharacterTable( "4.U4(3)" );;
    gap> List( [ "L3(4)", "2.L3(4)", "4_1.L3(4)", "4_2.L3(4)" ], name ->
    >          Length( PossibleClassFusions( CharacterTable( name ), t ) ) );
    [ 0, 0, 0, 4 ]

So the split extension 42.L3(4).23 of 42.L3(4) is a subgroup of the split extension 4.U4(3).23 of 4.U4(3), and only one of the two possible tables of 42.L3(4).23 admits a class fusion into the ATLAS table of 4.U3(4).23; the construction of the latter table is shown in Section 4.3.
    gap> t2:= CharacterTable( "4.U4(3).2_3" );;
    gap> List( poss, x -> Length( PossibleClassFusions( x.table, t2 ) ) );
    [ 0, 16, 0, 0 ]

I do not know a character theoretic argument that would disprove the existence of a group whose character table is the other candidate (or its isoclinic variant). For example, the table passes the tests from Section 4.12.
(It is straightforward to compute all extensions of 42.L3(4) by an automorphism of order two. The extensions with 34 conjugacy classes belong to the second candidate and its isoclinic variant.)
The correct table is the one that is contained in the GAP Character Table Library.
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[2].table,
    >                  lib ) );
    true
    gap> ConstructModularMGATables( tblMG, tblGA, lib );;

In the construction of the character table of 122.L3(4).23, the same ambiguity arises. We resolve it using the fact that 42.L3(4).23 occurs as a factor group, modulo the unique normal subgroup of order three.
    gap> tblMG := CharacterTable( "12_2.L3(4)" );;
    gap> tblG  := CharacterTable( "6.L3(4)" );;
    gap> tblGA := CharacterTable( "6.L3(4).2_3" );;
    gap> name  := "new12_2.L3(4).2_3";;
    gap> lib   := CharacterTable( "12_2.L3(4).2_3" );;
    gap> poss  := ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );;
    #E  4 possibilities for new12_2.L3(4).2_3
    gap> Length( poss );
    4
    gap> nsg:= ClassPositionsOfNormalSubgroups( poss[1].table );
    [ [ 1 ], [ 1, 5 ], [ 1, 7 ], [ 1, 4 .. 7 ], [ 1, 3 .. 7 ], [ 1 .. 7 ], 
      [ 1 .. 50 ], [ 1 .. 62 ] ]
    gap> List( nsg, x -> Sum( SizesConjugacyClasses( poss[1].table ){ x } ) );
    [ 1, 3, 2, 4, 6, 12, 241920, 483840 ]
    gap> factlib:= CharacterTable( "4_2.L3(4).2_3" );;
    gap> List( poss, x -> IsRecord( TransformingPermutationsCharacterTables(
    >                         x.table / [ 1, 5 ], factlib ) ) );
    [ false, true, false, false ]
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[2].table,
    >                  lib ) );
    true
    gap> ConstructModularMGATables( tblMG, tblGA, lib );;

4.6  The Character Table of (22 ×F4(2)):2 < B (March 2003)

The sporadic simple group B contains a maximal subgroup [N] of type (22 ×F4(2)):2, which is the normalizer of a 2C element [x] in B (see [CCN+85,p. 217]).
We will see below that the normal Klein four group V in [N] contains two 2A elements in B. The 2A centralizer in B, a group of the structure 2.2E6(2).2, contains maximal subgroups of the type 22 ×F4(2). So the two 2A type subgroups C1, C2 in V are conjugate in [N], and Z = 〈x 〉 is the centre of [N].
ctblcons11.png
We start with computing the class fusion of the 22 ×F4(2) type subgroup U of [N] into B; in order to speed this up, we first compute the class fusion of the F4(2) subgroup of U into B (which is unique), and use it and the stored embedding into U for prescribing an approximation of the desired class fusion. Additionally, we prescribe (without loss of generality) that the first involution class in V is mapped to the class 2C of B.
    gap> f42:= CharacterTable( "F4(2)" );;
    gap> v4:= CharacterTable( "2^2" );;
    gap> dp:= v4 * f42;
    CharacterTable( "V4xF4(2)" )
    gap> b:= CharacterTable( "B" );;
    gap> f42fusb:= PossibleClassFusions( f42, b );;
    gap> Length( f42fusb );
    1
    gap> f42fusdp:= GetFusionMap( f42, dp );;
    gap> comp:= CompositionMaps( f42fusb[1], InverseMap( f42fusdp ) );
    [ 1, 3, 3, 3, 5, 6, 6, 7, 9, 9, 9, 9, 14, 14, 13, 13, 10, 14, 14, 12, 14, 17,
      15, 18, 22, 22, 22, 22, 26, 26, 22, 22, 27, 27, 28, 31, 31, 39, 39, 36, 36,
      33, 33, 39, 39, 35, 41, 42, 47, 47, 49, 49, 49, 58, 58, 56, 56, 66, 66, 66,
      66, 58, 58, 66, 66, 69, 69, 60, 72, 72, 75, 79, 79, 81, 81, 85, 86, 83, 83,
      91, 91, 94, 94, 104, 104, 109, 109, 116, 116, 114, 114, 132, 132, 140, 140 ]
    gap> v4fusdp:= GetFusionMap( v4, dp );
    [ 1, 96 .. 286 ]
    gap> comp[ v4fusdp[2] ]:= 4;;
    gap> dpfusb:= PossibleClassFusions( dp, b, rec( fusionmap:= comp ) );;
    gap> Length( dpfusb );
    4
    gap> Set( List( dpfusb, x -> x{ v4fusdp } ) );
    [ [ 1, 4, 2, 2 ] ]

As announced above, we see that V contains two 2A involutions.
Set G = U / Z, M.G = U, and G.A = [N] / Z. The latter group is the direct product of F4(2).2 and a cyclic group of order 2. Next we compute the class fusion from G into G.A.
    gap> tblG:= dp / v4fusdp{ [ 1, 2 ] };;
    gap> tblMG:= dp;;
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> tblGA:= c2 * CharacterTable( "F4(2).2" );
    CharacterTable( "C2xF4(2).2" )
    gap> GfusGA:= PossibleClassFusions( tblG, tblGA );;
    gap> Length( GfusGA );
    4
    gap> Length( RepresentativesFusions( tblG, GfusGA, tblGA ) );
    1

In principle, we have to be careful which of these equivalent maps we choose, since the underlying symmetries may be broken in the central extension M.G → G, for which we choose the default factor fusion.
However, in this situation the fusion G into G.A is unique already up to table automorphisms of the table of G.A, so we are free to choose one map.
    gap> Length( RepresentativesFusions( Group( () ), GfusGA, tblGA ) );
    1
    gap> StoreFusion( tblG, GfusGA[1], tblGA );

The tables involved determine the character table of M.G.A ≅ [N] uniquely.
    gap> elms:= PossibleActionsForTypeMGA( tblMG, tblG, tblGA );;
    gap> Length( elms );
    1
    gap> poss:= PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA, elms[1],
    >               "(2^2xF4(2)):2" );;
    gap> Length( poss );
    1
    gap> tblMGA:= poss[1].table;;

Finally, we compare the table we constructed with the one that is contained in the GAP Character Table Library.
    gap> IsRecord( TransformingPermutationsCharacterTables( tblMGA,
    >                  CharacterTable( "(2^2xF4(2)):2" ) ) );
    true

4.7  The Character Table of 2.(S3 ×Fi22.2) ≤ 2.B (March 2003)

The sporadic simple group B contains a maximal subgroup [M] of type S3 ×Fi22.2. In order to compute the character table of its preimage M in the Schur cover 2.B, we first analyse the structure of M and then describe the construction of the character table from known character tables.
Let Z denote the centre of 2.B. We start with [M] = M/Z. Its class fusion into B is uniquely determined by the character tables.
    gap> s3:= CharacterTable( "Dihedral", 6 );;
    gap> fi222:= CharacterTable( "Fi22.2" );;
    gap> tblMbar:= s3 * fi222;;
    gap> b:= CharacterTable( "B" );;
    gap> Mbarfusb:= PossibleClassFusions( tblMbar, b );;
    gap> Length( Mbarfusb );
    1

The subgroup of type Fi22 lifts to the double cover 2.Fi22 (that is, a group that is not a direct product 2 ×Fi22) in 2.B since 2.B admits no class fusion from Fi22.
    gap> 2b:= CharacterTable( "2.B" );;
    gap> PossibleClassFusions( CharacterTable( "Fi22" ), 2b );
    [  ]

So the preimage of Fi22.2 is one of the two nonisomorphic but isoclinic groups of type 2.Fi22.2, and we have to decide which one really occurs. For that, we consider the subgroup of type 3 ×Fi22.2 in B, which is a 3A centralizer in B. Its preimage has the structure 3 ×2.Fi22.2 because the preimage of the central group of order 3 is a cyclic group of order 6 and thus contains a normal complement of the 2.Fi22 type subgroup. And a class fusion into 2.B is possible only from the direct product containing the 2.Fi22.2 group that is printed in the ATLAS.
    gap> c3:= CharacterTable( "Cyclic", 3 );;
    gap> 2fi222:= CharacterTable( "2.Fi22.2" );;
    gap> PossibleClassFusions( c3 * CharacterTableIsoclinic( 2fi222 ), 2b );
    [  ]

Next we note that the involutions in the normal subgroup [S] of type S3 in [M] lift to involutions in 2.B.
    gap> s3inMbar:= GetFusionMap( s3, tblMbar );
    [ 1, 113 .. 225 ]
    gap> s3inb:= Mbarfusb[1]{ s3inMbar };
    [ 1, 6, 2 ]
    gap> 2bfusb:= GetFusionMap( 2b, b );;
    gap> 2s3in2B:= InverseMap( 2bfusb ){ s3inb };
    [ [ 1, 2 ], [ 8, 9 ], 3 ]
    gap> CompositionMaps( OrdersClassRepresentatives( 2b ), 2s3in2B );
    [ [ 1, 2 ], [ 3, 6 ], 2 ]

Thus the preimage S of [S] contains elements of order 6 but no elements of order 4, which implies that S is a direct product 2 ×S3.
The two complements C1, C2 of Z in S are normal in the preimage N of [N] = S3 ×Fi22, which is thus of type S3 ×2.Fi22. However, they are conjugate under the action of 2.Fi22.2, as no class fusion from S3 ×2.Fi22.2 into 2.B is possible.
    gap> PossibleClassFusions( s3 * 2fi222, 2b );
    [  ]

(More specifically, the classes of element order 36 in 2.Fi22.2 have centralizer orders 36 and 72, so their centralizer orders in S3 ×2.Fi22.2 are 216 and 432; but the centralizers of order 36 elements in 2.B have centralizer order at most 216.)
Now let us see how the character table of M can be constructed.
Let Y denote the normal subgroup of order 3 in M, and U its centralizer in M, which has index 2 in M. Then the character table of M is determined by the tables of M/Y, U, U/Y ≅ 2.Fi22.2, and the action of M on the classes of U.
As for M/Y, consider the normal subgroup N = NM(C1) of index 2 in M. In particular, S/Y is central in N/Y but not in M/Y, so the character table of M/Y is determined by the tables of M/(YZ), N/Y ≅ 2 ×2.Fi22, N/(YZ) ≅ 2 ×Fi22, and the action of M/Y on the classes of N/Y.
Thus we proceed in two steps, starting with the computation of the character table of M/Y, for which we choose the name according to the structure 22.Fi22.2.
ctblcons12.png
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> 2fi22:= CharacterTable( "2.Fi22" );;
    gap> tblNmodY:= c2 * 2fi22;;
    gap> centre:= GetFusionMap( 2fi22, tblNmodY ){
    >                 ClassPositionsOfCentre( 2fi22 ) };
    [ 1, 2 ]
    gap> tblNmod6:= tblNmodY / centre;;
    gap> tblMmod6:= c2 * fi222;;
    gap> fus:= PossibleClassFusions( tblNmod6, tblMmod6 );;
    gap> Length( fus );
    1
    gap> StoreFusion( tblNmod6, fus[1], tblMmod6 );
    gap> elms:= PossibleActionsForTypeMGA( tblNmodY, tblNmod6, tblMmod6 );;
    gap> Length( elms );
    1
    gap> poss:= PossibleCharacterTablesOfTypeMGA( tblNmodY, tblNmod6, tblMmod6,
    >               elms[1], "2^2.Fi22.2" );;
    gap> Length( poss );
    1
    gap> tblMmodY:= poss[1].table;
    CharacterTable( "2^2.Fi22.2" )

So we found a unique solution for the character table of M/Y. Now we compute the table of M. For that, we have to specify the class fusion of U/Y into M/Y; it is unique up to table automorphisms of M/Y.
    gap> tblU:= c3 * 2fi222;;
    gap> tblUmodY:= tblU / GetFusionMap( c3, tblU );;
    gap> fus:= PossibleClassFusions( tblUmodY, tblMmodY );;
    gap> Length( RepresentativesFusions( Group( () ), fus, tblMmodY ) );
    1
    gap> StoreFusion( tblUmodY, fus[1], tblMmodY );
    gap> elms:= PossibleActionsForTypeMGA( tblU, tblUmodY, tblMmodY );;
    gap> Length( elms );
    1
    gap> poss:= PossibleCharacterTablesOfTypeMGA( tblU, tblUmodY, tblMmodY,
    >               elms[1], "(S3x2.Fi22).2" );;
    gap> Length( poss );
    1
    gap> tblM:= poss[1].table;
    CharacterTable( "(S3x2.Fi22).2" )
    gap> mfus2b:= PossibleClassFusions( tblM, 2b );;
    gap> Length( RepresentativesFusions( tblM, mfus2b, 2b ) );
    1

We did not construct M as a central extension of [M], so we verify that the tables fit together; note that this way we get also the class fusion from M onto [M].
    gap> Irr( tblM / ClassPositionsOfCentre( tblM ) ) = Irr( tblMbar );
    true

Finally, we compare the table we constructed with the one that is contained in the GAP Character Table Library.
    gap> IsRecord( TransformingPermutationsCharacterTables( tblM,
    >                  CharacterTable( "(S3x2.Fi22).2" ) ) );
    true

4.8  The Character Table of (2 ×2.Fi22):2 < Fi24 (November 2008)

The automorphism group Fi24 of the sporadic simple group Fi24 contains a maximal subgroup N of the type (2 ×2.Fi22):2, whose intersection with Fi24 is 2.Fi22.2 (see [CCN+85,p. 207]).
The normal Klein four group V in N contains two 2C elements in Fi24, because the 2C centralizer in Fi24, a group of the structure 2 ×Fi23, contains maximal subgroups of the type 2 ×2.Fi22, and so the two 2C type subgroups C1, C2 in V are conjugate in N, and Z = Z(N) is the centre of N ∩Fi24.
ctblcons100.png
With U = CN(C1), a group of the type 2 ×2.Fi22, we set G = U / Z, M.G = U, and G.A = N / Z. The latter group is the direct product of Fi22.2 and a cyclic group of order 2.
This is exactly the situation of the construction of the character table of the group that is called 22.Fi22.2 in Section 4.7, where this group occurs as "M/Y". Since the character table is uniquely determined by the input data, it is the table we are interested in here.
So all we have to do is to compute the class fusion from this table into that of Fi24.
    gap> fi24:= CharacterTable( "Fi24" );;
    gap> t:= CharacterTable( "2^2.Fi22.2" );;
    gap> fus:= PossibleClassFusions( t, fi24 );;
    gap> Length( fus );
    4
    gap> Length( RepresentativesFusions( t, fus, fi24 ) );
    1

(It should be noted that we did not need the character table of the 2.Fi22.2 type subgroup of N in the above construction, only the tables of 2.Fi22 and Fi22.2 were used.)
The fact that the character table of a factor of a subgroup of 2.B occurs as the character table of a subgroup of Fi24 is not a coincidence. In fact, the groups 3.Fi24 and 2.B are subgroups of the Monster group M, and the subgroup U = 2.(S3 ×Fi22.2) of 2.B normalizes an element of order three. The full normalizer of this element in M is 3.Fi24, which means that we have established U as a (maximal) subgroup of 3.Fi24. Note that we have constructed the character table of U in Section 4.7.
Let us compute the class fusion of U into 3.Fi24.
    gap> t:= CharacterTable( "(S3x2.Fi22).2" );;
    gap> 3fi24:= CharacterTable( "3.Fi24" );;                        
    gap> fus:= PossibleClassFusions( t, 3fi24 );;
    gap> Length( fus );
    16
    gap> Length( RepresentativesFusions( t, fus, 3fi24 ) );
    1
    gap> GetFusionMap( t, 3fi24 ) in fus; 
    true

Moreover, U turns out to be the full normalizer of a 6A element in M,
    gap> m:= CharacterTable( "M" );;
    gap> tfusm:= PossibleClassFusions( t, m );;
    gap> Length( tfusm );
    4
    gap> Length( RepresentativesFusions( t, tfusm, m ) );
    1
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >        x -> Sum( SizesConjugacyClasses( t ){ x } ) = 6 );
    [ [ 1, 2, 142, 143 ] ]
    gap> Set( List( tfusm, x -> x{ nsg[1] } ) );
    [ [ 1, 2, 4, 13 ] ]
    gap> OrdersClassRepresentatives( t ){ nsg[1] };
    [ 1, 2, 3, 6 ]
    gap> PowerMap( m, -1 )[13];
    13
    gap> Size( t ) = 2 * SizesCentralizers( m )[13];
    true

(Thus U is also the full normalizer of an element of order six in 2.B and in 3.Fi24.)

4.9  The Character Table of S3 ×2.U4(3).22 ≤ 2.Fi22 (September 2002)

The sporadic simple Fischer group Fi22 contains a maximal subgroup [M] of type S3 ×U4(3).22 (see [CCN+85,p. 163]). We claim that the preimage M of [M] in the central extension 2.Fi22 has the structure S3 ×2.U4(3).22, where the factor of type 2.U4(3).22 is the one printed in the ATLAS.
For that, we first note that the normal subgroup [S] of type S3 in [M] lifts to a group S which has the structure 2 ×S3. This follows from the fact that all involutions in Fi22 lift to involutions in 2.Fi22 or, equivalently, the central involution in 2.Fi22 is not a square.
ctblcons13.png
    gap> 2Fi22:= CharacterTable( "2.Fi22" );;
    gap> ClassPositionsOfCentre( 2Fi22 );
    [ 1, 2 ]
    gap> 2 in PowerMap( 2Fi22, 2 );
    false

Second, the normal subgroup [U] ≅ U4(3).22 of Fi22 lifts to a nonsplit extension U in 2.Fi22, since 2.Fi22 contains no U4(3) type subgroup. Furthermore, U is the 2.U4(3).22 type group printed in the ATLAS because the isoclinic variant does not admit a class fusion into 2.Fi22.
    gap> PossibleClassFusions( CharacterTable( "U4(3)" ), 2Fi22 );
    [  ]
    gap> tblU:= CharacterTable( "2.U4(3).2_2" );;
    gap> iso:= CharacterTableIsoclinic( tblU );
    CharacterTable( "Isoclinic(2.U4(3).2_2)" )
    gap> PossibleClassFusions( iso, 2Fi22 );                      
    [  ]

Now there are just two possibilities. Either the two S3 type subgroups in S are normal in M (and thus M is the direct product of any such S3 with the preimage of the U4(3).22 type subgroup), or they are conjugate in M.
Suppose we are in the latter situation, let z be a generator of the centre of 2.Fi22, and let τ, σ be an involution and an order three element respectively, in one of the S3 type subgroups.
Each element g ∈ U \U conjugates τ to an involution in the other S3 type subgroup of S, so g−1 τg = τσi z for some i ∈ { 0, 1, 2 }. Furthermore, it is possible to choose g as an involution.
    gap> derpos:= ClassPositionsOfDerivedSubgroup( tblU );;
    gap> outer:= Difference( [ 1 .. NrConjugacyClasses( tblU ) ], derpos );;
    gap> 2 in OrdersClassRepresentatives( tblU ){ outer };
    true

With this choice, (g τ)2 = τσi z τ = σ−i z holds, which means that (g τ)3 squares to z. As we have seen above, this is impossible, hence M is a direct product, as claimed.
The class fusion of M into 2.Fi22 is determined by the character tables, up to table automorphisms.
    gap> tblM:= CharacterTable( "Dihedral", 6 ) * tblU;;
    gap> fus:= PossibleClassFusions( tblM, 2Fi22 );;
    gap> Length( RepresentativesFusions( tblM, fus, 2Fi22 ) );
    1
    gap> IsRecord( TransformingPermutationsCharacterTables( tblM,
    >                  CharacterTable( "2.Fi22M8" ) ) );
    true

4.10  The Character Table of 4.HS.2 ≤ HN.2 (May 2002)

The maximal subgroup U of type 2.HS.2 in the sporadic simple group HN extends to a group N of structure 4.HS.2 in the automorphism group HN.2 of HN (see [CCN+85,p. 166]).
N is the normalizer of a 4D element g ∈ HN.2 \HN. The centralizer C of g is of type 4.HS, which is the central product of 2.HS and the cyclic group 〈g 〉 of order 4. We have Z = Z(N) = 〈g2 〉. Since U/Z ≅ HS.2 is a complement of 〈g 〉/ Z in N/Z, the factor group N/Z is a direct product of HS.2 and a cyclic group of order 2.
ctblcons14.png
Thus N has the structure 2.G.2, the normal subgroup 2.G being C, the factor group G.2 being 2 ×HS.2, and G being 2 ×HS. Each element in N \C inverts g, so N acts fixed point freely on the faithful irreducible characters of C. Hence we can use PossibleCharacterTablesOfTypeMGA for constructing the character table of N from the tables of C and N/Z and the action of N on the classes of C.
We start with the table of the central product C. It can be viewed as an isoclinic table of the direct product of 2.HS and a cyclic group of order 2, see 2.4.
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> tblC:= CharacterTableIsoclinic( CharacterTable( "2.HS" ) * c2 );;

The table of G is given as that of the factor group by the unique normal subgroup of C that consists of two conjugacy classes.
    gap> ord2:= Filtered( ClassPositionsOfNormalSubgroups( tblC ),
    >               x -> Length( x ) = 2 );
    [ [ 1, 3 ] ]
    gap> tblCbar:= tblC / ord2[1];;

Finally, we construct the table of the extension G.2 and the class fusion of G into this table (which is uniquely determined by the character tables).
    gap> tblNbar:= CharacterTable( "HS.2" ) * c2;;
    gap> fus:= PossibleClassFusions( tblCbar, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 
          21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 29, 30, 31, 32, 33, 34, 35, 36, 
          35, 36, 37, 38, 39, 40, 41, 42, 41, 42 ] ]
    gap> StoreFusion( tblCbar, fus[1], tblNbar );

Now we compute the table automorphisms of the table of C that are compatible with the extension N; we get two solutions.
    gap> elms:= PossibleActionsForTypeMGA( tblC, tblCbar, tblNbar );
    [ [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6, 8 ], [ 7 ], [ 9 ], [ 10 ], [ 11 ],
          [ 12, 14 ], [ 13 ], [ 15 ], [ 16, 18 ], [ 17 ], [ 19 ], [ 20 ], [ 21 ],
          [ 22 ], [ 23 ], [ 24, 26 ], [ 25 ], [ 27 ], [ 28, 30 ], [ 29 ], [ 31 ],
          [ 32, 34 ], [ 33 ], [ 35 ], [ 36, 38 ], [ 37 ], [ 39 ], [ 40, 42 ],
          [ 41 ], [ 43 ], [ 44, 46 ], [ 45 ], [ 47 ], [ 48, 50 ], [ 49 ],
          [ 51, 53 ], [ 52, 54 ], [ 55 ], [ 56, 58 ], [ 57 ], [ 59 ], [ 60 ],
          [ 61, 65 ], [ 62, 68 ], [ 63, 67 ], [ 64, 66 ], [ 69 ], [ 70, 72 ],
          [ 71 ], [ 73 ], [ 74, 76 ], [ 75 ], [ 77, 81 ], [ 78, 84 ], [ 79, 83 ],
          [ 80, 82 ] ],
      [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6, 8 ], [ 7 ], [ 9 ], [ 10 ], [ 11 ],
          [ 12, 14 ], [ 13 ], [ 15, 17 ], [ 16 ], [ 18 ], [ 19 ], [ 20 ], [ 21 ],
          [ 22 ], [ 23 ], [ 24, 26 ], [ 25 ], [ 27 ], [ 28, 30 ], [ 29 ], [ 31 ],
          [ 32, 34 ], [ 33 ], [ 35, 37 ], [ 36 ], [ 38 ], [ 39 ], [ 40, 42 ],
          [ 41 ], [ 43 ], [ 44, 46 ], [ 45 ], [ 47, 49 ], [ 48 ], [ 50 ],
          [ 51, 53 ], [ 52, 54 ], [ 55 ], [ 56, 58 ], [ 57 ], [ 59 ], [ 60 ],
          [ 61, 65 ], [ 62, 68 ], [ 63, 67 ], [ 64, 66 ], [ 69, 71 ], [ 70 ],
          [ 72 ], [ 73 ], [ 74, 76 ], [ 75 ], [ 77, 83 ], [ 78, 82 ], [ 79, 81 ],
          [ 80, 84 ] ] ]

We compute the possible character tables arising from these two actions.
    gap> poss:= List( elms, pi -> PossibleCharacterTablesOfTypeMGA(
    >                 tblC, tblCbar, tblNbar, pi, "4.HS.2" ) );;
    gap> List( poss, Length );
    [ 0, 2 ]

So one of the two table automorphisms turned out to be impossible; the reason is that the corresponding "character table" would not admit a 2-power map. (Alternatively, we could exclude this action on C by the fact that it is not compatible with the action of 2.HS.2 on its subgroup 2.HS, which occurs here as the restriction of the action of N on C to that of U on C ∩U.)
The other table automorphism leads to two possible character tables. This is not surprising since N contains a subgroup of type 2.HS.2, and the above setup does not determine which of the two isoclinism types of this group occurs. Let us look at the possible class fusions from these tables into that of HN.2:
    gap> result:= poss[2];;
    gap> hn2:= CharacterTable( "HN.2" );;
    gap> possfus:= List( result, r -> PossibleClassFusions( r.table, hn2 ) );;
    gap> List( possfus, Length );
    [ 32, 0 ]
    gap> RepresentativesFusions( result[1].table, possfus[1], hn2 );
    [ [ 1, 46, 2, 2, 47, 3, 7, 45, 4, 58, 13, 6, 46, 47, 6, 47, 7, 48, 10, 62, 
          20, 9, 63, 21, 12, 64, 24, 27, 49, 50, 13, 59, 14, 16, 70, 30, 18, 53, 
          52, 17, 54, 20, 65, 22, 36, 56, 26, 76, 39, 77, 28, 59, 58, 31, 78, 41, 
          34, 62, 35, 65, 2, 45, 3, 45, 6, 48, 7, 47, 17, 54, 13, 49, 13, 50, 14, 
          50, 18, 53, 18, 52, 21, 56, 25, 57, 27, 59, 30, 60, 44, 72, 34, 66, 35, 
          66, 41, 71 ] ]

Only one of the candidates admits an embedding, and the class fusion is unique up to table automorphisms. So we are done.
Finally, we compare the table we have constructed with the one that is contained in the GAP Character Table Library.
    gap> libtbl:= CharacterTable( "4.HS.2" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( result[1].table,
    >                  libtbl ) );
    true

(The following paragraphs have been added in May 2006.)
The Brauer tables of N = 2.G.2 can be constructed as in Section 4.3. Note that the Brauer tables of C = 2.G and of N / Z = G.2 are automatically available because the ordinary tables constructed above arose as a direct product and as an isoclinic table of a direct product, and the GAP Character Table Library contains the Brauer tables of the direct factors involved.
    gap> StoreFusion( tblC, result[1].MGfusMGA, result[1].table );
    gap> ForAll( Set( Factors( Size( result[1].table ) ) ),
    >            p -> IsRecord( TransformingPermutationsCharacterTables(
    >                     BrauerTableOfTypeMGA( tblC mod p, tblNbar mod p,
    >                         result[1].table ).table, libtbl mod p ) ) );
    true

Here it is advantageous that the Brauer table of C / Z = G is not needed in the construction, since GAP does not know how to compute the p-modular table of the ordinary table of G constructed above. Of course we have G ≅ 2 ×HS, and the p-modular table of HS is known, but in the construction of the table of G as a factor of the table of 2.G, the information is missing that the nonsolvable simple direct factor of 2.G corresponds to the library table of HS.

4.11  The Character Tables of 4.A6.23, 12.A6.23, and 4.L2(25).23

For the "broken box" cases in the ATLAS (see [CCN+85,p. xxiv]), the character tables can be constructed with the M.G.A construction method from Section 3.1.
The group N = 4.A6.23 (see [CCN+85,p. 5]) can be described as an upward extension of the normal subgroup C ≅ 4.A6 -which is a central product of U = 2.A6 and a cyclic group 〈g 〉 of order 4- by a cyclic group of order 2, such that the factor group of N by the central subgroup Z = 〈g2 〉 of order 2 is isomorphic to a subdirect product [N] of M10 = A6.23 and a cyclic group of order 4 and that N acts nontrivially on its normal subgroup 〈g 〉.
ctblcons15.png
Thus N has the structure 2.G.2, with 2.G = C and G.2 = [N]. These two groups are isoclinic variants of 2 ×2.A6 and of 2 ×M10, respectively. Each element in N \C inverts g, so it acts fixed point freely on the faithful irreducible characters of C. Hence we can use PossibleCharacterTablesOfTypeMGA for constructing the character table of N from the tables of C and N/Z and the action of N on the classes of C.
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> 2a6:= CharacterTable( "2.A6" );;
    gap> tblC:= CharacterTableIsoclinic( 2a6 * c2 );;
    gap> ord2:= Filtered( ClassPositionsOfNormalSubgroups( tblC ),
    >               x -> Length( x ) = 2 );
    [ [ 1, 3 ] ]
    gap> tblG:= tblC / ord2[1];;
    gap> tblNbar:= CharacterTableIsoclinic( CharacterTable( "A6.2_3" ) * c2 );;
    gap> fus:= PossibleClassFusions( tblG, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 5, 6, 7, 8, 9, 10, 9, 10 ] ]
    gap> StoreFusion( tblG, fus[1], tblNbar );
    gap> elms:= PossibleActionsForTypeMGA( tblC, tblG, tblNbar );
    [ [ [ 1 ], [ 2 ], [ 3 ], [ 4 ], [ 5 ], [ 6 ], [ 7, 11 ], [ 8, 12 ],
          [ 9, 13 ], [ 10, 14 ], [ 15, 17 ], [ 16, 18 ], [ 19, 23 ], [ 20, 24 ],
          [ 21, 25 ], [ 22, 26 ] ],
      [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6 ], [ 7, 11 ], [ 8, 14 ], [ 9, 13 ],
          [ 10, 12 ], [ 15 ], [ 16, 18 ], [ 17 ], [ 19, 23 ], [ 20, 26 ],
          [ 21, 25 ], [ 22, 24 ] ],
      [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6 ], [ 7, 11 ], [ 8, 14 ], [ 9, 13 ],
          [ 10, 12 ], [ 15, 17 ], [ 16 ], [ 18 ], [ 19, 23 ], [ 20, 26 ],
          [ 21, 25 ], [ 22, 24 ] ] ]
    gap> poss:= List( elms, pi -> PossibleCharacterTablesOfTypeMGA(
    >                 tblC, tblG, tblNbar, pi, "4.A6.2_3" ) );
    [ [  ], [  ], 
      [
          rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 9, 6, 9, 8, 7, 10, 11, 10,
                  12, 13, 14, 15, 16, 13, 16, 15, 14 ],
              table := CharacterTable( "4.A6.2_3" ) ) ] ]

So we get a unique solution. It coincides with the character table of 4.A6.23 that is stored in the GAP Character Table Library.
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[3][1].table,
    >                  CharacterTable( "4.A6.2_3" ) ) );
    true

Note that the first two candidates for the action lead to tables that do not admit a 2-power map. In fact the 2-power map of the character table of 4.A6.23 is not uniquely determined by the matrix of character values. However, the 2-power map is unique up to automorphisms of this matrix; the function PossibleCharacterTablesOfTypeMGA takes this into account, and returns only representatives, in this case one table.
The ATLAS states in [CCN+85,Section 6.7] that there is s group of the structure 22.A6.23 that is isoclinic with 4.A6.23. We construct also the character table of the 22.A6.23 type group with the M.G.A construction method from Section 3.1.
The group N = 22.A6.23 can be described as an upward extension of the normal subgroup C ≅ 2 ×2.A6 by a cyclic group of order 2, such that the factor group of N by the central subgroup Z of order 2 that is contained in U = C′ ≅ 2.A6 is isomorphic to a subdirect product [N] of M10 = A6.23 and a cyclic group of order 4 and that N acts nontrivially on the centre of C, which is a Klein four group.
ctblcons15.png
Thus N has the structure 2.G.2, with 2.G = C and G.2 = [N]. These latter group is an isoclinic variant of 2 ×M10, as in the construction of 4.A6.23. Each element in N \C swaps the two involutions in Z(C) \Z, so it acts fixed point freely on those irreducible characters of C whose kernels do not contain Z. Hence we can use PossibleCharacterTablesOfTypeMGA for constructing the character table of N from the tables of C and N/Z and the action of N on the classes of C.
    gap> tblC:= 2a6 * c2;;
    gap> z:= GetFusionMap( 2a6, tblC ){ ClassPositionsOfCentre( 2a6 ) };
    [ 1, 3 ]
    gap> tblG:= tblC / z;;
    gap> tblNbar:= CharacterTableIsoclinic( CharacterTable( "A6.2_3" ) * c2 );;
    gap> fus:= PossibleClassFusions( tblG, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 5, 6, 7, 8, 9, 10, 9, 10 ] ]
    gap> StoreFusion( tblG, fus[1], tblNbar );
    gap> elms:= PossibleActionsForTypeMGA( tblC, tblG, tblNbar );
    [ [ [ 1 ], [ 2 ], [ 3 ], [ 4 ], [ 5 ], [ 6 ], [ 7, 11 ], [ 8, 12 ], 
          [ 9, 13 ], [ 10, 14 ], [ 15, 17 ], [ 16, 18 ], [ 19, 23 ], [ 20, 24 ], 
          [ 21, 25 ], [ 22, 26 ] ], 
      [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6 ], [ 7, 11 ], [ 8, 14 ], [ 9, 13 ], 
          [ 10, 12 ], [ 15 ], [ 16, 18 ], [ 17 ], [ 19, 23 ], [ 20, 26 ], 
          [ 21, 25 ], [ 22, 24 ] ], 
      [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6 ], [ 7, 11 ], [ 8, 14 ], [ 9, 13 ], 
          [ 10, 12 ], [ 15, 17 ], [ 16 ], [ 18 ], [ 19, 23 ], [ 20, 26 ], 
          [ 21, 25 ], [ 22, 24 ] ] ]
    gap> poss:= List( elms, pi -> PossibleCharacterTablesOfTypeMGA(
    >                 tblC, tblG, tblNbar, pi, "2^2.A6.2_3" ) );
    [ [  ], [  ], 
      [ 
          rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 9, 6, 9, 8, 7, 10, 11, 10, 
                  12, 13, 14, 15, 16, 13, 16, 15, 14 ], 
              table := CharacterTable( "2^2.A6.2_3" ) ) ] ]

So we get a unique solution.
The group N = 12.A6.23 (see [CCN+85,p. 5]) can be described as an upward extension of the normal subgroup C ≅ 12.A6 -which is a central product of U = 6.A6 and a cyclic group 〈g 〉 of order 4- by a cyclic group of order 2, such that the factor group of N by the central subgroup Z = 〈g2 〉 of order 2 is isomorphic to a subdirect product [N] of 3.M10 = 3.A6.23 and a cyclic group of order 4 and that N acts nontrivially on its normal subgroup 〈g 〉.
Note that N has a central subgroup Y, say, of order 3, so the situation here differs from that for groups of the type 12.G.2 with G one of L3(4), U4(3), where the action on the normal subgroup of order three is nontrivial.
ctblcons16.png
Thus N has the structure 2.G.2, with 2.G = C and G.2 = [N]. These two groups are isoclinic variants of 2 ×6.A6 and of 2 ×3.M10, respectively. Each element in N \C inverts g, so it acts fixed point freely on the faithful irreducible characters of C. Hence we can use PossibleCharacterTablesOfTypeMGA for constructing the character table of N from the tables of C and N/Z and the action of N on the classes of C.
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> tblC:= CharacterTableIsoclinic( CharacterTable( "6.A6" ) * c2 );;
    gap> ord2:= Filtered( ClassPositionsOfNormalSubgroups( tblC ),
    >               x -> Length( x ) = 2 );
    [ [ 1, 7 ] ]
    gap> tblG:= tblC / ord2[1];;
    gap> tblNbar:= CharacterTableIsoclinic( CharacterTable( "3.A6.2_3" ) * c2 );;
    gap> fus:= PossibleClassFusions( tblG, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 13, 14, 15, 16, 17, 18, 
          19, 20, 21, 22, 23, 24, 25, 26, 21, 22, 23, 24, 25, 26 ], 
      [ 1, 2, 5, 6, 3, 4, 7, 8, 11, 12, 9, 10, 13, 14, 13, 14, 15, 16, 19, 20, 
          17, 18, 21, 22, 25, 26, 23, 24, 21, 22, 25, 26, 23, 24 ] ]
    gap> rep:= RepresentativesFusions( Group( () ), fus, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 13, 14, 15, 16, 17, 18, 
          19, 20, 21, 22, 23, 24, 25, 26, 21, 22, 23, 24, 25, 26 ] ]
    gap> StoreFusion( tblG, rep[1], tblNbar );
    gap> elms:= PossibleActionsForTypeMGA( tblC, tblG, tblNbar );
    [ [ [ 1 ], [ 2 ], [ 3 ], [ 4 ], [ 5 ], [ 6 ], [ 7 ], [ 8 ], [ 9 ], [ 10 ], 
          [ 11 ], [ 12 ], [ 13 ], [ 14 ], [ 15 ], [ 16 ], [ 17 ], [ 18 ], 
          [ 19, 23 ], [ 20, 24 ], [ 21, 25 ], [ 22, 26 ], [ 27, 33 ], [ 28, 34 ], 
          [ 29, 35 ], [ 30, 36 ], [ 31, 37 ], [ 32, 38 ], [ 39, 51 ], [ 40, 52 ], 
          [ 41, 53 ], [ 42, 54 ], [ 43, 55 ], [ 44, 56 ], [ 45, 57 ], [ 46, 58 ], 
          [ 47, 59 ], [ 48, 60 ], [ 49, 61 ], [ 50, 62 ] ], 
      [ [ 1 ], [ 2, 8 ], [ 3 ], [ 4, 10 ], [ 5 ], [ 6, 12 ], [ 7 ], [ 9 ], 
          [ 11 ], [ 13 ], [ 14 ], [ 15 ], [ 16 ], [ 17 ], [ 18 ], [ 19, 23 ], 
          [ 20, 26 ], [ 21, 25 ], [ 22, 24 ], [ 27 ], [ 28, 34 ], [ 29 ], 
          [ 30, 36 ], [ 31 ], [ 32, 38 ], [ 33 ], [ 35 ], [ 37 ], [ 39, 51 ], 
          [ 40, 58 ], [ 41, 53 ], [ 42, 60 ], [ 43, 55 ], [ 44, 62 ], [ 45, 57 ], 
          [ 46, 52 ], [ 47, 59 ], [ 48, 54 ], [ 49, 61 ], [ 50, 56 ] ], 
      [ [ 1 ], [ 2, 8 ], [ 3 ], [ 4, 10 ], [ 5 ], [ 6, 12 ], [ 7 ], [ 9 ], 
          [ 11 ], [ 13 ], [ 14 ], [ 15 ], [ 16 ], [ 17 ], [ 18 ], [ 19, 23 ], 
          [ 20, 26 ], [ 21, 25 ], [ 22, 24 ], [ 27, 33 ], [ 28 ], [ 29, 35 ], 
          [ 30 ], [ 31, 37 ], [ 32 ], [ 34 ], [ 36 ], [ 38 ], [ 39, 51 ], 
          [ 40, 58 ], [ 41, 53 ], [ 42, 60 ], [ 43, 55 ], [ 44, 62 ], [ 45, 57 ], 
          [ 46, 52 ], [ 47, 59 ], [ 48, 54 ], [ 49, 61 ], [ 50, 56 ] ] ]
    gap> poss:= List( elms, pi -> PossibleCharacterTablesOfTypeMGA(
    >                 tblC, tblG, tblNbar, pi, "12.A6.2_3" ) );
    [ [  ], [  ], 
      [
          rec( MGfusMGA := [ 1, 2, 3, 4, 5, 6, 7, 2, 8, 4, 9, 6, 10, 11, 12, 13, 14,
                  15, 16, 17, 18, 19, 16, 19, 18, 17, 20, 21, 22, 23, 24, 25, 20,
                  26, 22, 27, 24, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39,
                  40, 29, 36, 31, 38, 33, 40, 35, 30, 37, 32, 39, 34 ],
              table := CharacterTable( "12.A6.2_3" ) ) ] ]

So we get again a unique solution. It coincides with the character table that is stored in the GAP Character Table Library.
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[3][1].table,
    >                  CharacterTable( "12.A6.2_3" ) ) );
    true

The construction of the character table of 4.L2(25).23 is analogous to that of the table of 4.A6.23. We get a unique table that coincides with the table in the GAP library.
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> tblC:= CharacterTableIsoclinic( CharacterTable( "2.L2(25)" ) * c2 );;
    gap> ord2:= Filtered( ClassPositionsOfNormalSubgroups( tblC ),
    >               x -> Length( x ) = 2 );
    [ [ 1, 3 ] ]
    gap> tblG:= tblC / ord2[1];;
    gap> tblNbar:= CharacterTableIsoclinic( CharacterTable( "L2(25).2_3" ) * c2 );;
    gap> fus:= PossibleClassFusions( tblG, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 9, 10, 11, 12, 13, 14, 13, 14, 15, 16, 15, 
          16, 17, 18, 17, 18, 19, 20, 19, 20 ], 
      [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 9, 10, 11, 12, 13, 14, 13, 14, 17, 18, 17, 
          18, 19, 20, 19, 20, 15, 16, 15, 16 ], 
      [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 9, 10, 11, 12, 13, 14, 13, 14, 19, 20, 19, 
          20, 15, 16, 15, 16, 17, 18, 17, 18 ] ]
    gap> rep:= RepresentativesFusions( Group( () ), fus, tblNbar );
    [ [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 9, 10, 11, 12, 13, 14, 13, 14, 15, 16, 15, 
          16, 17, 18, 17, 18, 19, 20, 19, 20 ] ]
    gap> StoreFusion( tblG, rep[1], tblNbar );
    gap> elms:= PossibleActionsForTypeMGA( tblC, tblG, tblNbar );
    [ [ [ 1 ], [ 2 ], [ 3 ], [ 4 ], [ 5 ], [ 6 ], [ 7 ], [ 8 ], [ 9 ], [ 10 ], 
          [ 11, 13 ], [ 12, 14 ], [ 15, 19 ], [ 16, 20 ], [ 17, 21 ], [ 18, 22 ], 
          [ 23, 25 ], [ 24, 26 ], [ 27, 33 ], [ 28, 34 ], [ 29, 31 ], [ 30, 32 ], 
          [ 35, 39 ], [ 36, 40 ], [ 37, 41 ], [ 38, 42 ], [ 43, 47 ], [ 44, 48 ], 
          [ 45, 49 ], [ 46, 50 ], [ 51, 55 ], [ 52, 56 ], [ 53, 57 ], [ 54, 58 ] ]
        , [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6 ], [ 7 ], [ 8, 10 ], [ 9 ], 
          [ 11 ], [ 12, 14 ], [ 13 ], [ 15, 19 ], [ 16, 22 ], [ 17, 21 ], 
          [ 18, 20 ], [ 23, 25 ], [ 24 ], [ 26 ], [ 27, 31 ], [ 28, 34 ], 
          [ 29, 33 ], [ 30, 32 ], [ 35, 39 ], [ 36, 42 ], [ 37, 41 ], [ 38, 40 ], 
          [ 43, 47 ], [ 44, 50 ], [ 45, 49 ], [ 46, 48 ], [ 51, 55 ], [ 52, 58 ], 
          [ 53, 57 ], [ 54, 56 ] ], 
      [ [ 1 ], [ 2, 4 ], [ 3 ], [ 5 ], [ 6 ], [ 7 ], [ 8, 10 ], [ 9 ], 
          [ 11, 13 ], [ 12 ], [ 14 ], [ 15, 19 ], [ 16, 22 ], [ 17, 21 ], 
          [ 18, 20 ], [ 23, 25 ], [ 24 ], [ 26 ], [ 27, 33 ], [ 28, 32 ], 
          [ 29, 31 ], [ 30, 34 ], [ 35, 39 ], [ 36, 42 ], [ 37, 41 ], [ 38, 40 ], 
          [ 43, 47 ], [ 44, 50 ], [ 45, 49 ], [ 46, 48 ], [ 51, 55 ], [ 52, 58 ], 
          [ 53, 57 ], [ 54, 56 ] ] ]
    gap> poss:= List( elms, pi -> PossibleCharacterTablesOfTypeMGA(
    >                 tblC, tblG, tblNbar, pi, "4.L2(25).2_3" ) );
    [ [  ], [  ], 
      [
          rec( MGfusMGA := [ 1, 2, 3, 2, 4, 5, 6, 7, 8, 7, 9, 10, 9, 11, 12, 13, 14,
                  15, 12, 15, 14, 13, 16, 17, 16, 18, 19, 20, 21, 22, 21, 20, 19,
                  22, 23, 24, 25, 26, 23, 26, 25, 24, 27, 28, 29, 30, 27, 30, 29,
                  28, 31, 32, 33, 34, 31, 34, 33, 32 ],
              table := CharacterTable( "4.L2(25).2_3" ) ) ] ]
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[3][1].table,
    >                  CharacterTable( "4.L2(25).2_3" ) ) );
    true

4.12   Pseudo Character Tables of the Type M.G.A (May 2004)

With the construction method for character tables of groups of the type M.G.A, one can construct tables that have many properties of character tables but that are not character tables of groups, cf. [Gag86]. For example, the group 3.A6.23 has a central subgroup of order 3, so it is not of the type M.G.A with fixed-point free action on the faithful characters of M.G.
However, if we apply the "M.G.A construction" to the groups M.G = 3.A6, G = A6, and G.A = A6.23 then we get a (in this case unique) result.
    gap> tblMG := CharacterTable( "3.A6" );;
    gap> tblG  := CharacterTable( "A6" );;
    gap> tblGA := CharacterTable( "A6.2_3" );;
    gap> elms:= PossibleActionsForTypeMGA( tblMG, tblG, tblGA );  
    [ [ [ 1 ], [ 2, 3 ], [ 4 ], [ 5, 6 ], [ 7, 8 ], [ 9 ], [ 10, 11 ],
          [ 12, 15 ], [ 13, 17 ], [ 14, 16 ] ] ]
    gap> poss:= PossibleCharacterTablesOfTypeMGA(                  
    >                 tblMG, tblG, tblGA, elms[1], "pseudo" );    
    [ rec( MGfusMGA := [ 1, 2, 2, 3, 4, 4, 5, 5, 6, 7, 7, 8, 9, 10, 8, 10, 9 ],
          table := CharacterTable( "pseudo" ) ) ]

Such a table automatically satisfies the orthogonality relations, and the tensor product of two "irreducible characters" of which at least one is a row from G.A decomposes into a sum of the "irreducible characters", where the coefficients are nonnegative integers.
In this example, any tensor product decomposes with nonnegative integral coefficients, n-th symmetrizations of "irreducible characters" decompose, for n ≤ 5, and the "class multiplication coefficients" are nonnegative integers.
    gap> pseudo:= poss[1].table;
    CharacterTable( "pseudo" )
    gap> Display( pseudo );
    pseudo
    
          2  4   3  4  3  .  3   2  .   .   .  2  3  3
          3  3   3  1  1  2  1   1  1   1   1  .  .  .
          5  1   1  .  .  .  .   .  1   1   1  .  .  .
    
            1a  3a 2a 6a 3b 4a 12a 5a 15a 15b 4b 8a 8b
         2P 1a  3a 1a 3a 3b 2a  6a 5a 15a 15b 2a 4a 4a
         3P 1a  1a 2a 2a 1a 4a  4a 5a  5a  5a 4b 8a 8b
         5P 1a  3a 2a 6a 3b 4a 12a 1a  3a  3a 4b 8b 8a
    
    X.1      1   1  1  1  1  1   1  1   1   1  1  1  1
    X.2      1   1  1  1  1  1   1  1   1   1 -1 -1 -1
    X.3     10  10  2  2  1 -2  -2  .   .   .  .  .  .
    X.4     16  16  .  . -2  .   .  1   1   1  .  .  .
    X.5      9   9  1  1  .  1   1 -1  -1  -1  1 -1 -1
    X.6      9   9  1  1  .  1   1 -1  -1  -1 -1  1  1
    X.7     10  10 -2 -2  1  .   .  .   .   .  .  B -B
    X.8     10  10 -2 -2  1  .   .  .   .   .  . -B  B
    X.9      6  -3 -2  1  .  2  -1  1   A  /A  .  .  .
    X.10     6  -3 -2  1  .  2  -1  1  /A   A  .  .  .
    X.11    12  -6  4 -2  .  .   .  2  -1  -1  .  .  .
    X.12    18  -9  2 -1  .  2  -1 -2   1   1  .  .  .
    X.13    30 -15 -2  1  . -2   1  .   .   .  .  .  .
    
    A = -E(15)-E(15)^2-E(15)^4-E(15)^8
      = (-1-Sqrt(-15))/2 = -1-b15
    B = E(8)+E(8)^3
      = Sqrt(-2) = i2
    gap> IsInternallyConsistent( pseudo );
    true
    gap> irr:= Irr( pseudo );;
    gap> test:= Concatenation( List( [ 2 .. 5 ],
    >               n -> Symmetrizations( pseudo, irr, n ) ) );;
    gap> Append( test, Set( Tensored( irr, irr ) ) );
    gap> fail in Decomposition( irr, test, "nonnegative" );
    false
    gap> if ForAny( Tuples( [ 1 .. NrConjugacyClasses( pseudo ) ], 3 ),        
    >      t -> not ClassMultiplicationCoefficient( pseudo, t[1], t[2], t[3] )   
    >               in NonnegativeIntegers ) then                           
    >      Error( "contradiction" );
    > fi;

I do not know a character-theoretic argument for showing that this table is not the character table of a group, but we can use the following group-theoretic argument. Suppose that the group G, say, has the above character table. Then G has a unique composition series with factors of the orders 3, 360, and 2, respectively. Let N denote the normal subgroup of order 3 in G. The factor group F = G/N is an automorphic extension of A6, and according to [CCN+85,p. 4] it is isomorphic with M10 = A6.23 and has Sylow 3 normalizers of the structure 32 : Q8. Since the Sylow 3 subgroup of G is a self-centralizing nonabelian group of order 33 and of exponent 3, the Sylow 3 normalizers in G have the structure 31+2+ : Q8, but the Q8 type subgroups of Aut( 31+2+ ) act trivially on the centre of 31+2+, contrary to the situation in the above table.
In general, this construction need not produce tables for which all symmetrizations of irreducible characters decompose properly. For example, applying PossibleCharacterTablesOfTypeMGA to the case M.G = 3.L3(4) and G.A = L3(4).21 does not yield a table because the function suppresses tables that do not admit p-th power maps, for prime divisors p of the order of M.G.A, and in this case no compatible 2-power map exists.
    gap> tblMG := CharacterTable( "3.L3(4)" );;
    gap> tblG  := CharacterTable( "L3(4)" );;
    gap> tblGA := CharacterTable( "L3(4).2_1" );;
    gap> elms:= PossibleActionsForTypeMGA( tblMG, tblG, tblGA );
    [ [ [ 1 ], [ 2, 3 ], [ 4 ], [ 5, 6 ], [ 7 ], [ 8 ], [ 9, 10 ], [ 11 ], 
          [ 12, 13 ], [ 14 ], [ 15, 16 ], [ 17, 20 ], [ 18, 22 ], [ 19, 21 ], 
          [ 23, 26 ], [ 24, 28 ], [ 25, 27 ] ] ]
    gap> PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA, elms[1], "?" );
    [  ]

Also, it may happen that already PossibleActionsForTypeMGA returns an empty list. Examples are M.G = 31.U4(3), G.A = U4(3).22 and M.G = 32.U4(3), G.A = U4(3).23.
    gap> tblG  := CharacterTable( "U4(3)" );;
    gap> tblMG := CharacterTable( "3_1.U4(3)" );;
    gap> tblGA := CharacterTable( "U4(3).2_2" );;
    gap> PossibleActionsForTypeMGA( tblMG, tblG, tblGA );
    [  ]
    gap> tblMG:= CharacterTable( "3_2.U4(3)" );;
    gap> tblGA:= CharacterTable( "U4(3).2_3" );;
    gap> PossibleActionsForTypeMGA( tblMG, tblG, tblGA );
    [  ]

4.13  Some Extra-ordinary p-Modular Tables of the Type M.G.A (September 2005)

For a group M.G.A in the sense of Section 3.1 such that not all ordinary irreducible characters χ have the property that M is contained in the kernel of χ or χ is induced from M.G, it may happen that there are primes p such that all irreducible p-modular characters have this property. This happens if and only if the preimages in M.G.A of each p-regular conjugacy class in G.A \G form one conjugacy class.
The following function can be used to decide whether this situation applies to a character table in the GAP Character Table Library; here we assume that for the library table of a group with the structure M.G.A, the class fusions from M.G and to G.A are stored.
    gap> FindExtraordinaryCase:= function( tblMGA )
    >    local result, der, nsg, tblMGAclasses, orders, tblMG,
    >          tblMGfustblMGA, tblMGclasses, pos, M, Mimg, tblMGAfustblGA, tblGA,
    >          outer, inv, filt, other, primes, p;
    >    result:= [];
    >    der:= ClassPositionsOfDerivedSubgroup( tblMGA );
    >    nsg:= ClassPositionsOfNormalSubgroups( tblMGA );
    >    tblMGAclasses:= SizesConjugacyClasses( tblMGA );
    >    orders:= OrdersClassRepresentatives( tblMGA );
    >    if Length( der ) < NrConjugacyClasses( tblMGA ) then
    >      # Look for tables of normal subgroups of the form $M.G$.
    >      for tblMG in Filtered( List( NamesOfFusionSources( tblMGA ),
    >                                   CharacterTable ), x -> x <> fail ) do
    >        tblMGfustblMGA:= GetFusionMap( tblMG, tblMGA );
    >        tblMGclasses:= SizesConjugacyClasses( tblMG );
    >        pos:= Position( nsg, Set( tblMGfustblMGA ) );
    >        if pos <> fail and
    >           Size( tblMG ) = Sum( tblMGAclasses{ nsg[ pos ] } ) then
    >          # Look for normal subgroups of the form $M$.
    >          for M in Difference( ClassPositionsOfNormalSubgroups( tblMG ),
    >                       [ [ 1 ], [ 1 .. NrConjugacyClasses( tblMG ) ] ] ) do
    >            Mimg:= Set( tblMGfustblMGA{ M } );
    >            if Sum( tblMGAclasses{ Mimg } ) = Sum( tblMGclasses{ M } ) then
    >              tblMGAfustblGA:= First( ComputedClassFusions( tblMGA ),
    >                  r -> ClassPositionsOfKernel( r.map ) = Mimg );
    >              if tblMGAfustblGA <> fail then
    >                tblGA:= CharacterTable( tblMGAfustblGA.name );
    >                tblMGAfustblGA:= tblMGAfustblGA.map;
    >                outer:= Difference( [ 1 .. NrConjugacyClasses( tblGA ) ],
    >                    CompositionMaps( tblMGAfustblGA, tblMGfustblMGA ) );
    >                inv:= InverseMap( tblMGAfustblGA ){ outer };
    >                filt:= Flat( Filtered( inv, IsList ) );
    >                if not IsEmpty( filt ) then
    >                  other:= Filtered( inv, IsInt );
    >                  primes:= Filtered( Set( Factors( Size( tblMGA ) ) ),
    >                     p -> ForAll( orders{ filt }, x -> x mod p = 0 )
    >                          and ForAny( orders{ other }, x -> x mod p <> 0 ) );
    >                  for p in primes do
    >                    Add( result, [ Identifier( tblMG ),
    >                                   Identifier( tblMGA ),
    >                                   Identifier( tblGA ), p ] );
    >                  od;
    >                fi;
    >              fi;
    >            fi;
    >          od;
    >        fi;
    >      od;
    >    fi;
    >    return result;
    > end;;

Let us list the tables which are found by this function.
    gap> cases:= [];;
    gap> for name in AllCharacterTableNames() do
    >      Append( cases, FindExtraordinaryCase( CharacterTable( name ) ) );
    >    od;
    gap> for i in Set( cases ) do
    >      Print( i, "\n" ); 
    >    od;
    [ "2.A6", "2.A6.2_1", "A6.2_1", 3 ]
    [ "2.Fi22", "2.Fi22.2", "Fi22.2", 3 ]
    [ "2.L2(25)", "2.L2(25).2_2", "L2(25).2_2", 5 ]
    [ "2.L2(49)", "2.L2(49).2_2", "L2(49).2_2", 7 ]
    [ "2.L2(81)", "2.L2(81).2_1", "L2(81).2_1", 3 ]
    [ "2.L2(81)", "2.L2(81).4_1", "L2(81).4_1", 3 ]
    [ "2.L2(81).2_1", "2.L2(81).4_1", "L2(81).4_1", 3 ]
    [ "2.L4(3)", "2.L4(3).2_2", "L4(3).2_2", 3 ]
    [ "2.L4(3)", "2.L4(3).2_3", "L4(3).2_3", 3 ]
    [ "2.U4(3).2_1", "2.U4(3).(2^2)_{12*2*}", "U4(3).(2^2)_{122}", 3 ]
    [ "2.U4(3).2_1", "2.U4(3).(2^2)_{122}", "U4(3).(2^2)_{122}", 3 ]
    [ "2.U4(3).2_1", "2.U4(3).(2^2)_{13*3*}", "U4(3).(2^2)_{133}", 3 ]
    [ "2.U4(3).2_1", "2.U4(3).(2^2)_{133}", "U4(3).(2^2)_{133}", 3 ]
    [ "3.U3(8)", "3.U3(8).3_1", "U3(8).3_1", 2 ]
    [ "3.U3(8)", "3.U3(8).6", "U3(8).6", 2 ]
    [ "3.U3(8)", "3.U3(8).6", "U3(8).6", 3 ]
    [ "3.U3(8).2", "3.U3(8).6", "U3(8).6", 2 ]
    [ "5^(1+2):8:4", "2.HS.2N5", "HS.2N5", 5 ]
    [ "6.A6", "6.A6.2_1", "3.A6.2_1", 3 ]
    [ "6.A6", "6.A6.2_1", "A6.2_1", 3 ]
    [ "6.Fi22", "6.Fi22.2", "3.Fi22.2", 3 ]
    [ "6.Fi22", "6.Fi22.2", "Fi22.2", 3 ]
    [ "Isoclinic(2.U4(3).2_1)", "2.U4(3).(2^2)_{1*2*2}", "U4(3).(2^2)_{122}", 3 ]
    [ "Isoclinic(2.U4(3).2_1)", "2.U4(3).(2^2)_{1*3*3}", "U4(3).(2^2)_{133}", 3 ]

The smallest example in this list is 2.A6.21, the double cover of the symmetric group on six points. The 3-modular table of this group looks as follows.
    gap> Display( CharacterTable( "2.A6.2_1" ) mod 3 );
    2.A6.2_1mod3
    
         2  5   5  4  3  1   1  4  4  3
         3  2   2  .  .  .   .  1  1  .
         5  1   1  .  .  1   1  .  .  .
    
           1a  2a 4a 8a 5a 10a 2b 4b 8b
        2P 1a  1a 2a 4a 5a  5a 1a 2a 4a
        3P 1a  2a 4a 8a 5a 10a 2b 4b 8b
        5P 1a  2a 4a 8a 1a  2a 2b 4b 8b
    
    X.1     1   1  1  1  1   1  1  1  1
    X.2     1   1  1  1  1   1 -1 -1 -1
    X.3     6   6 -2  2  1   1  .  .  .
    X.4     4   4  . -2 -1  -1  2 -2  .
    X.5     4   4  . -2 -1  -1 -2  2  .
    X.6     9   9  1  1 -1  -1  3  3 -1
    X.7     9   9  1  1 -1  -1 -3 -3  1
    X.8     4  -4  .  . -1   1  .  .  .
    X.9    12 -12  .  .  2  -2  .  .  .

We see that the two faithful irreducible characters vanish on the three classes outside 2.A6.
For the groups in the above list, the function BrauerTableOfTypeMGA can be used to construct the p-modular tables of M.G.A from the tables of M.G and G.A, for the given special primes p. The computations can be performed as follows.
    gap> for input in cases do
    >      p:= input[4];
    >      modtblMG:=  CharacterTable( input[1] ) mod p;
    >      ordtblMGA:= CharacterTable( input[2] );
    >      modtblGA:=  CharacterTable( input[3] ) mod p;
    >      name:= Concatenation( Identifier( ordtblMGA ), " mod ", String(p) );
    >      if ForAll( [ modtblMG, modtblGA ], IsCharacterTable ) then
    >        poss:= BrauerTableOfTypeMGA( modtblMG, modtblGA, ordtblMGA );
    >        modlib:= ordtblMGA mod p;
    >        if IsCharacterTable( modlib ) then
    >          trans:= TransformingPermutationsCharacterTables( poss.table,
    >                      modlib );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " differ\n" );
    >          fi;
    >        else
    >          Print( "#I  no library table for ", name, "\n" );
    >        fi;
    >      else
    >        Print( "#I  not all input tables for ", name, " available\n" );
    >      fi;
    >    od;
    #I  not all input tables for 2.L2(49).2_2 mod 7 available
    #I  not all input tables for 2.L2(81).2_1 mod 3 available
    #I  not all input tables for 2.L2(81).4_1 mod 3 available
    #I  not all input tables for 2.L2(81).4_1 mod 3 available

The examples 2.A6.21, 2.L2(25).22, and 2.L2(49).22 belong to the infinite series of semiliniear groups ΣL(2,p2), for odd primes p. All groups in this series have the property that all faithful irreducible characters vanish on the p-regular classes outside SL(2,p2). (Cf. Section 2.5 for another property of the groups in this series.)

5  Examples for the Type G.S3

5.1  Small Examples

The symmetric group S4 on four points has the form G.S3 where G is the Klein four group V4, G.2 is the dihedral group D8 of order 8, and G.3 is the alternating group A4. The trivial character of A4 extends twofold to S4, in the same way as the trivial character of V4 extends to the dihedral group. The nontrivial linear characters of A4 induce irreducibly to S4. The irreducible degree three character of A4 is induced from any of the three nontrivial linear characters of V4, it extends to S4 in the same way as the unique constituent of the restriction to V4 that is invariant in the chosen D8 extends to D8.
    gap> c2:= CharacterTable( "Cyclic", 2 );;
    gap> t:= c2 * c2;;
    gap> tC:= CharacterTable( "Dihedral", 8 );;
    gap> tK:= CharacterTable( "Alternating", 4 );;
    gap> tfustC:= PossibleClassFusions( t, tC );
    [ [ 1, 3, 4, 4 ], [ 1, 3, 5, 5 ], [ 1, 4, 3, 4 ], [ 1, 4, 4, 3 ], 
      [ 1, 5, 3, 5 ], [ 1, 5, 5, 3 ] ]
    gap> StoreFusion( t, tfustC[1], tC );
    gap> tfustK:= PossibleClassFusions( t, tK );
    [ [ 1, 2, 2, 2 ] ]
    gap> StoreFusion( t, tfustK[1], tK );
    gap> elms:= PossibleActionsForTypeGS3( t, tC, tK );
    [ (3,4) ]
    gap> new:= CharacterTableOfTypeGS3( t, tC, tK, elms[1], "S4" );
    rec( table := CharacterTable( "S4" ), tblCfustblKC := [ 1, 4, 2, 2, 5 ], 
      tblKfustblKC := [ 1, 2, 3, 3 ] )
    gap> Display( new.table );
    S4
    
         2  3  3  .  2  2
         3  1  .  1  .  .
    
           1a 2a 3a 4a 2b
        2P 1a 1a 3a 2a 1a
        3P 1a 2a 1a 4a 2b
    
    X.1     1  1  1  1  1
    X.2     1  1  1 -1 -1
    X.3     3 -1  .  1 -1
    X.4     3 -1  . -1  1
    X.5     2  2 -1  .  .

The case e > 1 occurs in the following example. We choose G the cyclic group of order two, G.C the cyclic group of order six, G.K the quaternion group of order eight, and construct the character table of G.F = SL2(3), with F ≅ A4.
We get three extensions of the trivial character of G.K to G.F, a degree three character induced from the nontrivial linear characters of G.K, and three extensions of the irreducible degree 2 character of G.K.
    gap> t:= CharacterTable( "Cyclic", 2 );;
    gap> tC:= CharacterTable( "Cyclic", 6 );;
    gap> tK:= CharacterTable( "Quaternionic", 8 );;
    gap> tfustC:= PossibleClassFusions( t, tC );
    [ [ 1, 4 ] ]
    gap> StoreFusion( t, tfustC[1], tC );
    gap> tfustK:= PossibleClassFusions( t, tK );
    [ [ 1, 3 ] ]
    gap> StoreFusion( t, tfustK[1], tK );
    gap> elms:= PossibleActionsForTypeGS3( t, tC, tK );
    [ (2,5,4) ]
    gap> new:= CharacterTableOfTypeGS3( t, tC, tK, elms[1], "SL(2,3)" );
    rec( table := CharacterTable( "SL(2,3)" ), 
      tblCfustblKC := [ 1, 4, 5, 3, 6, 7 ], tblKfustblKC := [ 1, 2, 3, 2, 2 ] )
    gap> Display( new.table );
    SL(2,3)
    
         2  3  2  3  1   1   1  1
         3  1  .  1  1   1   1  1
    
           1a 4a 2a 6a  3a  3b 6b
        2P 1a 2a 1a 3a  3b  3a 3b
        3P 1a 4a 2a 2a  1a  1a 2a
    
    X.1     1  1  1  1   1   1  1
    X.2     1  1  1  A  /A   A /A
    X.3     1  1  1 /A   A  /A  A
    X.4     3 -1  3  .   .   .  .
    X.5     2  . -2 /A  -A -/A  A
    X.6     2  . -2  1  -1  -1  1
    X.7     2  . -2  A -/A  -A /A
    
    A = E(3)
      = (-1+Sqrt(-3))/2 = b3

5.2   ATLAS Tables of the Type G.S3

We demonstrate the construction of all those ordinary and modular character tables in the GAP Character Table Library that are of the type G.S3 where G is a simple group or a central extension of a simple group whose character table is contained in the ATLAS. Here is the list of Identifier values needed for accessing the input tables and the known library tables corresponding to the output.
    gap> listGS3:= [
    > [ "U3(5)",      "U3(5).2",      "U3(5).3",      "U3(5).S3"        ],
    > [ "3.U3(5)",    "3.U3(5).2",    "3.U3(5).3",    "3.U3(5).S3"      ],
    > [ "L3(4)",      "L3(4).2_2",    "L3(4).3",      "L3(4).3.2_2"     ],
    > [ "L3(4)",      "L3(4).2_3",    "L3(4).3",      "L3(4).3.2_3"     ],
    > [ "3.L3(4)",    "3.L3(4).2_2",  "3.L3(4).3",    "3.L3(4).3.2_2"   ],
    > [ "2^2.L3(4)",  "2^2.L3(4).2_2","2^2.L3(4).3",  "2^2.L3(4).3.2_2" ],
    > [ "2^2.L3(4)",  "2^2.L3(4).2_3","2^2.L3(4).3",  "2^2.L3(4).3.2_3" ],
    > [ "U6(2)",      "U6(2).2",      "U6(2).3",      "U6(2).3.2"       ],
    > [ "3.U6(2)",    "3.U6(2).2",    "3.U6(2).3",    "3.U6(2).3.2"     ],
    > [ "2^2.U6(2)",  "2^2.U6(2).2",  "2^2.U6(2).3",  "2^2.U6(2).3.2"   ],
    > [ "O8+(2)",     "O8+(2).2",     "O8+(2).3",     "O8+(2).3.2"      ],
    > [ "2^2.O8+(2)", "2^2.O8+(2).2", "2^2.O8+(2).3", "2^2.O8+(2).3.2"  ],
    > [ "L3(7)",      "L3(7).2",      "L3(7).3",      "L3(7).S3"        ],
    > [ "3.L3(7)",    "3.L3(7).2",    "3.L3(7).3",    "3.L3(7).S3"      ],
    > [ "U3(8)",      "U3(8).2",      "U3(8).3_2",    "U3(8).S3"        ],
    > [ "3.U3(8)",    "3.U3(8).2",    "3.U3(8).3_2",  "3.U3(8).S3"      ],
    > [ "U3(11)",     "U3(11).2",     "U3(11).3",     "U3(11).S3"       ],
    > [ "3.U3(11)",   "3.U3(11).2",   "3.U3(11).3",   "3.U3(11).S3"     ],
    > [ "O8+(3)",     "O8+(3).2_2",   "O8+(3).3",     "O8+(3).S3"       ],
    > [ "2E6(2)",     "2E6(2).2",     "2E6(2).3",     "2E6(2).S3"       ],
    > [ "2^2.2E6(2)", "2^2.2E6(2).2", "2^2.2E6(2).3", "2^2.2E6(2).S3"   ],
    > ];;

(For G one of L3(4), U6(2), O8+(2), and 2E6(2), the tables of 22.G, 22.G.2, and 22.G.3 can be constructed with the methods described in Section 3.4 and Section 3.1, respectively.)
Analogously, the automorphism groups of L3(4) and O8+(3) have factor groups isomorphic with S3; in these cases, we choose G = L3(4).21 and G = O8+(3).22111, respectively.
    gap> Append( listGS3, [
    > [ "L3(4).2_1",          "L3(4).2^2",     "L3(4).6",     "L3(4).D12"     ],
    > [ "2^2.L3(4).2_1",      "2^2.L3(4).2^2", "2^2.L3(4).6", "2^2.L3(4).D12" ],
    > [ "O8+(3).(2^2)_{111}", "O8+(3).D8",     "O8+(3).A4",   "O8+(3).S4"     ],
    > ] );

In all these cases, the required table automorphism of G.3 is uniquely determined. We first compute the ordinary character table of G.S3 and then the p-modular tables, for all prime divisors p of the order of G such that the GAP Character Table Library contains the necessary p-modular input tables.
In each case, we compare the computed character tables with the ones in the GAP Character Table Library. Note that in order to avoid conflicts of the class fusions that arise in the construction with the class fusions that are already stored on the library tables, we choose identifiers for the result tables that are different from the identifiers of the library tables.
    gap> ProcessGS3Example:= function( t, tC, tK, identifier, pi )
    >    local tF, lib, trans, p, tmodp, tCmodp, tKmodp, modtF;
    > 
    >    tF:= CharacterTableOfTypeGS3( t, tC, tK, pi,
    >             Concatenation( identifier, "new" ) );
    >    lib:= CharacterTable( identifier );
    >    if lib <> fail then
    >      trans:= TransformingPermutationsCharacterTables( tF.table, lib );
    >      if not IsRecord( trans ) then
    >        Print( "#E  computed table and library table for `", identifier,
    >               "' differ\n" );
    >      fi;
    >    else
    >      Print( "#I  no library table for `", identifier, "'\n" );
    >    fi;
    >    StoreFusion( tC, tF.tblCfustblKC, tF.table );
    >    StoreFusion( tK, tF.tblKfustblKC, tF.table );
    >    for p in Set( Factors( Size( t ) ) ) do
    >      tmodp := t  mod p;
    >      tCmodp:= tC mod p;
    >      tKmodp:= tK mod p;
    >      if IsCharacterTable( tmodp ) and
    >         IsCharacterTable( tCmodp ) and
    >         IsCharacterTable( tKmodp ) then
    >        modtF:= CharacterTableOfTypeGS3( tmodp, tCmodp, tKmodp,
    >                    tF.table,
    >                    Concatenation(  identifier, "mod", String( p ) ) );
    >        if   Length( Irr( modtF.table ) ) <>
    >             Length( Irr( modtF.table )[1] ) then
    >          Print( "#E  nonsquare result table for `",
    >                 identifier, " mod ", p, "'\n" );
    >        elif lib <> fail and IsCharacterTable( lib mod p ) then
    >          trans:= TransformingPermutationsCharacterTables( modtF.table,
    >                                                           lib mod p );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for `",
    >                   identifier, " mod ", p, "' differ\n" );
    >          fi;
    >        else
    >          Print( "#I  no library table for `", identifier, " mod ",
    >                 p, "'\n" );
    >        fi;
    >      else
    >        Print( "#I  not all inputs available for `", identifier,
    >               " mod ", p, "'\n" );
    >      fi;
    >    od;
    > end;;

Now we call the function for the examples in the list.
    gap> for input in listGS3 do
    >      t := CharacterTable( input[1] );
    >      tC:= CharacterTable( input[2] );
    >      tK:= CharacterTable( input[3] );
    >      identifier:= input[4];
    >      elms:= PossibleActionsForTypeGS3( t, tC, tK );
    >      if Length( elms ) = 1 then
    >        ProcessGS3Example( t, tC, tK, identifier, elms[1] );
    >      else
    >        Print( "#I  ", Length( elms ), " actions possible for `",
    >               identifier, "'\n" );
    >      fi;
    >    od;
    #I  not all inputs available for `O8+(3).S3 mod 3'
    #I  not all inputs available for `2E6(2).S3 mod 2'
    #I  not all inputs available for `2E6(2).S3 mod 3'
    #I  not all inputs available for `2E6(2).S3 mod 5'
    #I  not all inputs available for `2E6(2).S3 mod 7'
    #I  not all inputs available for `2E6(2).S3 mod 11'
    #I  not all inputs available for `2E6(2).S3 mod 13'
    #I  not all inputs available for `2E6(2).S3 mod 17'
    #I  not all inputs available for `2E6(2).S3 mod 19'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 2'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 3'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 5'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 7'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 11'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 13'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 17'
    #I  not all inputs available for `2^2.2E6(2).S3 mod 19'
    #I  not all inputs available for `O8+(3).S4 mod 3'

Also the ordinary character table of the automorphic extension of the simple ATLAS group O8+(3) by A4 can be constructed with the same approach. Here we get four possible permutations, which lead to essentially the same character table.
    gap> input:= [ "O8+(3)", "O8+(3).3", "O8+(3).(2^2)_{111}", "O8+(3).A4" ];;
    gap> t := CharacterTable( input[1] );;
    gap> tC:= CharacterTable( input[2] );;
    gap> tK:= CharacterTable( input[3] );;
    gap> identifier:= input[4];;
    gap> elms:= PossibleActionsForTypeGS3( t, tC, tK );;
    gap> Length( elms );
    4
    gap> differ:= MovedPoints( Group( List( elms, x -> x / elms[1] ) ) );;
    gap> List( elms, x -> RestrictedPerm( x, differ ) );
    [ (118,216,169)(119,217,170)(120,218,167)(121,219,168), 
      (118,216,170)(119,217,169)(120,219,168)(121,218,167), 
      (118,217,169)(119,216,170)(120,218,168)(121,219,167), 
      (118,217,170)(119,216,169)(120,219,167)(121,218,168) ]
    gap> poss:= List( elms, pi -> CharacterTableOfTypeGS3( t, tC, tK, pi,
    >             Concatenation( identifier, "new" ) ) );;
    gap> lib:= CharacterTable( identifier );;
    gap> ForAll( poss, r -> IsRecord(
    >        TransformingPermutationsCharacterTables( r.table, lib ) ) );
    true

Also the construction of the p-modular tables of O8+(3).A4 works.
    gap> ProcessGS3Example( t, tC, tK, identifier, elms[1] );
    #I  not all inputs available for `O8+(3).A4 mod 3'

6  Examples for the Type G.22

6.1  The Character Table of A6.22

As the first example, we consider the automorphism group Aut( A6 ) ≅ A6.22 of the alternating group A6 on six points.
In this case, the triple of actions on the subgroups A6.2i is uniquely determined by the condition on the number of conjugacy classes in Section 3.3.
    gap> tblG:= CharacterTable( "A6" );;
    gap> tblsG2:= List( [ "A6.2_1", "A6.2_2", "A6.2_3" ], CharacterTable );;
    gap> List( tblsG2, NrConjugacyClasses );
    [ 11, 11, 8 ]
    gap> possact:= List( tblsG2, x -> Filtered( Elements( 
    >        AutomorphismsOfTable( x ) ), y -> Order( y ) <= 2 ) );
    [ [ (), (3,4)(7,8)(10,11) ], [ (), (8,9), (5,6)(10,11), (5,6)(8,9)(10,11) ], 
      [ (), (7,8) ] ]

Note that n1 = n2 implies f1 = f2, and n1 − n3 = 3 implies f1 − f3 = 2, so we get f1 = 3 and f3 = 1, and A6.22 has 2 ·11 − 3 ·3 = 2 ·8 − 3 ·1 = 13 classes.
(The compatibility on the classes inside A6 yields only that the classes 3 and 4 of A6.21 ≅ S6 must be fused in A6.22, as well as the classes 5 and 6 of A6.22PGL(2,9).)
    gap> List( tblsG2, x -> GetFusionMap( tblG, x ) );
    [ [ 1, 2, 3, 4, 5, 6, 6 ], [ 1, 2, 3, 3, 4, 5, 6 ], [ 1, 2, 3, 3, 4, 5, 5 ] ]

These arguments are used by the GAP function PossibleActionsForTypeGV4, which returns the list of all possible triples of permutations such that the i-th permutation describes the action of A6.22 on the classes of A6.2i.
    gap> acts:= PossibleActionsForTypeGV4( tblG, tblsG2 );    
    [ [ (3,4)(7,8)(10,11), (5,6)(8,9)(10,11), (7,8) ] ]

For the given actions, the GAP function PossibleCharacterTablesOfTypeGV4 then computes the possibilities for the character table of A6.22; in this case, the result is unique.
    gap> poss:= PossibleCharacterTablesOfTypeGV4( tblG, tblsG2, acts[1], "A6.2^2" );
    [ rec(
          G2fusGV4 := [ [ 1, 2, 3, 3, 4, 5, 6, 6, 7, 8, 8 ],
              [ 1, 2, 3, 4, 5, 5, 9, 10, 10, 11, 11 ],
              [ 1, 2, 3, 4, 5, 12, 13, 13 ] ],
          table := CharacterTable( "A6.2^2" ) ) ]
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[1].table,
    >                  CharacterTable( "A6.2^2" ) ) );
    true

Finally, possible p-modular tables can be computed from the p-modular input tables and the ordinary table of A6.22; here we show this for p = 3.
    gap> PossibleCharacterTablesOfTypeGV4( tblG mod 3,
    >        List( tblsG2, t -> t mod 3 ), poss[1].table );
    [ rec(
          G2fusGV4 := [ [ 1, 2, 3, 4, 5, 5, 6 ], [ 1, 2, 3, 4, 4, 7, 8, 8, 9, 9 ],
              [ 1, 2, 3, 4, 10, 11, 11 ] ], table := BrauerTable( "A6.2^2", 3 ) )
     ]

6.2   ATLAS Tables of the Type G.22 - Easy Cases

We demonstrate the construction of all those ordinary and modular character tables in the GAP Character Table Library that are of the type G.22 where G is a simple group or a central extension of a simple group whose character table is contained in the ATLAS. Here is the list of Identifier values needed for accessing the input tables and the result tables.
(The construction of the character table of O8+(3).22111 is more involved and will be described in Section 6.10. The construction of the character tables of groups of the type 2.L3(4).22 and 6.L3(4).22 is described in the sections 6.4 and 6.5, respectively. The construction of the character tables of groups of the type 2.U4(3).22 is described in Section 6.6.)
    gap> listGV4:= [
    > [ "A6",      "A6.2_1",      "A6.2_2",      "A6.2_3",      "A6.2^2"      ],
    > [ "3.A6",    "3.A6.2_1",    "3.A6.2_2",    "3.A6.2_3",    "3.A6.2^2"    ],
    > [ "L2(25)",  "L2(25).2_1",  "L2(25).2_2",  "L2(25).2_3",  "L2(25).2^2"  ],
    > [ "L3(4)",   "L3(4).2_1",   "L3(4).2_2",   "L3(4).2_3",   "L3(4).2^2"   ],
    > [ "2^2.L3(4)", "2^2.L3(4).2_1", "2^2.L3(4).2_2", "2^2.L3(4).2_3",
    >                                                         "2^2.L3(4).2^2" ],
    > [ "3.L3(4)", "3.L3(4).2_1", "3.L3(4).2_2", "3.L3(4).2_3", "3.L3(4).2^2" ],
    > [ "U4(3)",   "U4(3).2_1",   "U4(3).2_2",   "U4(3).2_2'",
    >                                                     "U4(3).(2^2)_{122}" ],
    > [ "U4(3)",   "U4(3).2_1",   "U4(3).2_3",   "U4(3).2_3'",
    >                                                     "U4(3).(2^2)_{133}" ],
    > [ "3_1.U4(3)", "3_1.U4(3).2_1", "3_1.U4(3).2_2", "3_1.U4(3).2_2'",
    >                                                 "3_1.U4(3).(2^2)_{122}" ],
    > [ "3_2.U4(3)", "3_2.U4(3).2_1", "3_2.U4(3).2_3", "3_2.U4(3).2_3'",
    >                                                 "3_2.U4(3).(2^2)_{133}" ],
    > [ "L2(49)",  "L2(49).2_1",  "L2(49).2_2",  "L2(49).2_3",  "L2(49).2^2"  ],
    > [ "L2(81)",  "L2(81).2_1",  "L2(81).2_2",  "L2(81).2_3",  "L2(81).2^2"  ],
    > [ "L3(9)",   "L3(9).2_1",   "L3(9).2_2",   "L3(9).2_3",   "L3(9).2^2"   ],
    > [ "O8+(3)",  "O8+(3).2_1",  "O8+(3).2_2",  "O8+(3).2_2'",
    >                                                    "O8+(3).(2^2)_{122}" ],
    > [ "O8-(3)",  "O8-(3).2_1",  "O8-(3).2_2",  "O8-(3).2_3",  "O8-(3).2^2"  ],
    > ];;

Analogously, the automorphism groups L3(4).D12 of L3(4) and U4(3).D8 of U4(3), and the subgroup O8+(3).D8 of the automorphism group O8+(3).S4 have factor groups that are isomorphic with 22; in these cases, we choose G = L3(4).3, G = U4(3).21, and G = O8+(3).21, respectively.
Also the group 22.L3(4).D12 has a factor group isomorphic with 22. Note that the character tables of L3(4).D12 and 22.L3(4).D12 have been constructed already in Section 5.2.
The automorphism groups of L4(4) and U4(5) have the structure L4(4).22 and U4(5).22, respectively; their tables are contained in the GAP Character Table Library but not in the ATLAS.
    gap> Append( listGV4, [
    > [ "L3(4).3", "L3(4).6",     "L3(4).3.2_2", "L3(4).3.2_3", "L3(4).D12"   ],
    > [ "2^2.L3(4).3", "2^2.L3(4).6", "2^2.L3(4).3.2_2", "2^2.L3(4).3.2_3",
    >                                                         "2^2.L3(4).D12" ],
    > [ "U4(3).2_1", "U4(3).4", "U4(3).(2^2)_{122}", "U4(3).(2^2)_{133}",
    >                                                              "U4(3).D8" ],
    > [ "O8+(3).2_1", "O8+(3).(2^2)_{111}", "O8+(3).(2^2)_{122}", "O8+(3).4",
    >                                                             "O8+(3).D8" ],
    > [ "L4(4)",   "L4(4).2_1",   "L4(4).2_2",   "L4(4).2_3",   "L4(4).2^2"   ],
    > [ "U4(5)",   "U4(5).2_1",   "U4(5).2_2",   "U4(5).2_3",   "U4(5).2^2"   ],
    > ] );

Now we proceed in two steps, the computation of the possible ordinary character tables from the ordinary tables of the relevant subgroups, and then the computation of the Brauer tables from the Brauer tables of the relevant subgroups and from the ordinary table of the group.
The following function first computes the possible triples of actions on the subgroups G.2i, using the function PossibleActionsForTypeGV4. Then the union of the candidate tables for these actions is computed, this list is returned in the end. and representatives of classes of permutation equivalent candidates are inspected further with consistency checks. If there is a unique solution up to permutation equivalence, this table is compared with the one that is contained in the GAP Character Table Library.
    gap> ConstructOrdinaryGV4Table:= function( tblG, tblsG2, name, lib )
    >      local acts, nam, poss, reps, i, trans;
    > 
    >      # Compute the possible actions for the ordinary tables.
    >      acts:= PossibleActionsForTypeGV4( tblG, tblsG2 );
    >      # Compute the possible ordinary tables for the given actions.
    >      nam:= Concatenation( "new", name );
    >      poss:= Concatenation( List( acts, triple -> 
    >          PossibleCharacterTablesOfTypeGV4( tblG, tblsG2, triple, nam ) ) );
    >      # Test the possibilities for permutation equivalence.
    >      reps:= RepresentativesCharacterTables( poss );
    >      if 1 < Length( reps ) then
    >        Print( "#I  ", name, ": ", Length( reps ),
    >               " equivalence classes\n" );
    >      elif Length( reps ) = 0 then
    >        Print( "#E  ", name, ": no solution\n" );
    >      else
    >        # Compare the computed table with the library table.
    >        if not IsCharacterTable( lib ) then
    >          Print( "#I  no library table for ", name, "\n" );
    >          PrintToLib( name, poss[1].table );
    >          for i in [ 1 .. 3 ] do
    >            Print( LibraryFusion( tblsG2[i],
    >                       rec( name:= name, map:= poss[1].G2fusGV4[i] ) ) );
    >          od;
    >        else
    >          trans:= TransformingPermutationsCharacterTables( poss[1].table,
    >                      lib );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " differ\n" );
    >          fi;
    >          # Compare the computed fusions with the stored ones.
    >          if List( poss[1].G2fusGV4, x -> OnTuples( x, trans.columns ) )
    >                 <> List( tblsG2, x -> GetFusionMap( x, lib ) ) then
    >            Print( "#E  computed and stored fusions for ", name,
    >                   " differ\n" );
    >          fi;
    >        fi;
    >      fi;
    >      return poss;
    >    end;;

The following function computes, for all those prime divisors p of the group order in question such that the p-modular Brauer tables of the subgroups G.2i are available, the possible p-modular Brauer tables. If the solution is unique up to permutation equivalence, it is compared with the table that is contained in the GAP Character Table Library.
It may happen (even in the case that the ordinary character table is unique up to permutation equivalence) that some candidates for the ordinary character table are excluded due to information provided by some p-modular table. In this case, a message is printed, and the ordinary character table from the GAP Character Table Library is checked again under the additional restrictions.
    gap> ConstructModularGV4Tables:= function( tblG, tblsG2, ordposs,
    >                                          ordlibtblGV4 )
    >      local name, modposs, primes, checkordinary, i, record, p, tmodp,
    >            t2modp, poss, modlib, trans, reps;
    > 
    >      if not IsCharacterTable( ordlibtblGV4 ) then
    >        Print( "#I  no ordinary library table ...\n" );
    >        return [];
    >      fi;
    >      name:= Identifier( ordlibtblGV4 );
    >      modposs:= [];
    >      primes:= Set( Factors( Size( tblG ) ) );
    >      ordposs:= ShallowCopy( ordposs );
    >      checkordinary:= false;
    >      for i in [ 1 .. Length( ordposs ) ] do
    >        modposs[i]:= [];
    >        record:= ordposs[i];
    >        for p in primes do
    >          tmodp := tblG  mod p;
    >          t2modp:= List( tblsG2, t2 -> t2 mod p );
    >          if IsCharacterTable( tmodp ) and
    >             ForAll( t2modp, IsCharacterTable ) then
    >            poss:= PossibleCharacterTablesOfTypeGV4( tmodp, t2modp,
    >                       record.table, record.G2fusGV4 );
    >            poss:= RepresentativesCharacterTables( poss );
    >            if   Length( poss ) = 0 then
    >              Print( "#I  excluded cand. ", i, " (out of ",
    >                     Length( ordposs ), ") for ", name, " by ", p,
    >                     "-mod. table\n" );
    >              Unbind( ordposs[i] );
    >              Unbind( modposs[i] );
    >              checkordinary:= true;
    >              break;
    >            elif Length( poss ) = 1 then
    >              # Compare the computed table with the library table.
    >              modlib:= ordlibtblGV4 mod p;
    >              if IsCharacterTable( modlib ) then
    >                trans:= TransformingPermutationsCharacterTables(
    >                            poss[1].table, modlib );
    >                if not IsRecord( trans ) then
    >                  Print( "#E  computed table and library table for ",
    >                         name, " mod ", p, " differ\n" );
    >                fi;
    >              else
    >                Print( "#I  no library table for ",
    >                       name, " mod ", p, "\n" );
    >                PrintToLib( name, poss[1].table );
    >              fi;
    >            else
    >              Print( "#I  ", name, " mod ", p, ": ", Length( poss ),
    >                     " equivalence classes\n" );
    >            fi;
    >            Add( modposs[i], poss );
    >          else
    >            Print( "#I  not all input tables for ", name, " mod ", p,
    >                   " available\n" );
    >            primes:= Difference( primes, [ p ] );
    >          fi;
    >        od;
    >      od;
    >      if checkordinary then
    >        # Test whether the ordinary table is admissible.
    >        ordposs:= Compacted( ordposs );
    >        modposs:= Compacted( modposs );
    >        reps:= RepresentativesCharacterTables( ordposs );
    >        if 1 < Length( reps ) then
    >          Print( "#I  ", name, ": ", Length( reps ),
    >                 " equivalence classes (ord. table)\n" );
    >        elif Length( reps ) = 0 then
    >          Print( "#E  ", name, ": no solution (ord. table)\n" );
    >        else
    >          # Compare the computed table with the library table.
    >          trans:= TransformingPermutationsCharacterTables(
    >                      ordposs[1].table, ordlibtblGV4 );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " differ\n" );
    >          fi;
    >          # Compare the computed fusions with the stored ones.
    >          if List( ordposs[1].G2fusGV4, x -> OnTuples( x, trans.columns ) )
    >               <> List( tblsG2, x -> GetFusionMap( x, ordlibtblGV4 ) ) then
    >            Print( "#E  computed and stored fusions for ", name,
    >                   " differ\n" );
    >          fi;
    >        fi;
    >      fi;
    >      return rec( ordinary:= ordposs, modular:= modposs );
    >    end;;

Finally, here is the loop over the list of tables.
    gap> for input in listGV4 do
    >      tblG   := CharacterTable( input[1] );
    >      tblsG2 := List( input{ [ 2 .. 4 ] }, CharacterTable );
    >      lib    := CharacterTable( input[5] );
    >      poss   := ConstructOrdinaryGV4Table( tblG, tblsG2, input[5], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >    od;
    #I  excluded cand. 2 (out of 2) for L3(4).2^2 by 3-mod. table
    #I  excluded cand. 2 (out of 8) for 2^2.L3(4).2^2 by 7-mod. table
    #I  excluded cand. 3 (out of 8) for 2^2.L3(4).2^2 by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 2^2.L3(4).2^2 by 5-mod. table
    #I  excluded cand. 5 (out of 8) for 2^2.L3(4).2^2 by 5-mod. table
    #I  excluded cand. 6 (out of 8) for 2^2.L3(4).2^2 by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 2^2.L3(4).2^2 by 7-mod. table
    #I  excluded cand. 2 (out of 2) for 3.L3(4).2^2 by 3-mod. table
    #I  not all input tables for L2(49).2^2 mod 7 available
    #I  not all input tables for L2(81).2^2 mod 3 available
    #I  excluded cand. 2 (out of 2) for L3(9).2^2 by 7-mod. table
    #I  not all input tables for O8+(3).(2^2)_{122} mod 3 available
    #I  not all input tables for O8-(3).2^2 mod 2 available
    #I  not all input tables for O8-(3).2^2 mod 3 available
    #I  not all input tables for O8-(3).2^2 mod 5 available
    #I  not all input tables for O8-(3).2^2 mod 7 available
    #I  not all input tables for O8-(3).2^2 mod 13 available
    #I  not all input tables for O8-(3).2^2 mod 41 available
    #I  excluded cand. 2 (out of 2) for L3(4).D12 by 3-mod. table
    #I  excluded cand. 2 (out of 2) for 2^2.L3(4).D12 by 7-mod. table
    #I  not all input tables for O8+(3).D8 mod 3 available
    #I  not all input tables for L4(4).2^2 mod 3 available
    #I  not all input tables for L4(4).2^2 mod 5 available
    #I  not all input tables for L4(4).2^2 mod 7 available
    #I  not all input tables for L4(4).2^2 mod 17 available
    #I  not all input tables for U4(5).2^2 mod 2 available
    #I  not all input tables for U4(5).2^2 mod 3 available
    #I  not all input tables for U4(5).2^2 mod 5 available
    #I  not all input tables for U4(5).2^2 mod 7 available
    #I  not all input tables for U4(5).2^2 mod 13 available

The groups 3.A6.22, 3.L3(4).22, and 32.U4(3).(22)133 have also the structure M.G.A, with M.G equal to 3.A6.23, 3.L3(4).21, and 32.U4(3).23, respectively, and G.A equal to A6.22, L3(4).22, and U4(3).(22)133, respectively (see Section 4.3).
Similarly, the group L3(4).D12 has also the structure G.S3, with G = L3(4).21, G.2 = L3(4).22, and G.3 = L3(4).6, respectively (see Section 5.2).

6.3  The Character Table of S4(9).22 (September 2011)

The available functions yield two possibilities for the ordinary character table of S4(9).22.
    gap> tblG:= CharacterTable( "S4(9)" );;
    gap> tblsG2:= List( [ "S4(9).2_1", "S4(9).2_2", "S4(9).2_3" ],
    >                   CharacterTable );;
    gap> lib:= CharacterTable( "S4(9).2^2" );;
    gap> poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, "newS4(9).2^2", lib );;
    #I  newS4(9).2^2: 2 equivalence classes
    gap> poss:= RepresentativesCharacterTables( poss );;

The two candidates differ w. r. t. the action of S4(9).22 on the classes of element order 80 in S4(9).22. In the two possible tables, each element of order 80 is conjugate to its third power or to its 13-th power, respectively.
    gap> order80:= PositionsProperty( OrdersClassRepresentatives( tblsG2[2] ),
    >                  x -> x = 80 );
    [ 98, 99, 100, 101, 102, 103, 104, 105 ]
    gap> List( poss, r -> r.G2fusGV4[2]{ order80 } );
    [ [ 77, 77, 78, 79, 80, 78, 79, 80 ], [ 77, 78, 79, 79, 77, 80, 80, 78 ] ]
    gap> PowerMap( tblsG2[2], 3 ){ order80 };
    [ 99, 98, 103, 104, 105, 100, 101, 102 ]
    gap> PowerMap( tblsG2[2], 13 ){ order80 };
    [ 102, 105, 101, 100, 98, 104, 103, 99 ]

We claim that the first candidate is the correct one. For that, first note that S4(9).22 is the extension of the simple group by a diagonal automorphism. (An easy way to see this is that for any subgroup of S4(9) isomorphic with S2(81) ≅ L2(81), the extension by a diagonal automorphism contains elements of order 80 -this group is isomorphic with PGL(2,81)- and only S4(9).22 contains elements of order 80.)
    gap> List( tblsG2, x -> 80 in OrdersClassRepresentatives( x ) );
    [ false, true, false ]

Now the field automorphism of S4(9).22 maps each element x of order 80 in S4(9).22 to a conjugate of x3.
    gap> tbl:= poss[1].table;;
    gap> IsRecord( TransformingPermutationsCharacterTables( tbl, lib ) );
    true

6.4  The Character Tables of Groups of the Type 2.L3(4).22 (June 2010)

The outer automorphism group of the group L3(4) is a dihedral group of order 12; its Sylow 2-subgroups are Klein four groups, so there is a unique almost simple group H of the type L3(4).22, up to isomorphism. In this section, we construct the character tables of the double covers of this group with the approach from Section 3.3.
The group H has three subgroups of index two, of the types L3(4).21, L3(4).22, and L3(4).23, respectively. So any double cover of H contains one subgroup of each of the types 2.L3(4).21, 2.L3(4).22, and 2.L3(4).23, and there are two isoclinic variants of each of these group to consider, see Section 2.5. So we start with eight different inputs for the construction of the character tables of double covers.
    gap> names:= List( [ 1 .. 3 ],
    >                  i -> Concatenation( "2.L3(4).2_", String( i ) ) );;
    gap> tbls:= List( names, CharacterTable );
    [ CharacterTable( "2.L3(4).2_1" ), CharacterTable( "2.L3(4).2_2" ), 
      CharacterTable( "2.L3(4).2_3" ) ]
    gap> isos:= List( names, nam -> CharacterTable( Concatenation( nam, "*" ) ) );
    [ CharacterTable( "Isoclinic(2.L3(4).2_1)" ), 
      CharacterTable( "Isoclinic(2.L3(4).2_2)" ), 
      CharacterTable( "Isoclinic(2.L3(4).2_3)" ) ]
    gap> inputs:= [
    > [ tbls[1], tbls[2], tbls[3], "2.L3(4).(2^2)_{123}" ],
    > [ tbls[1], isos[2], tbls[3], "2.L3(4).(2^2)_{12*3}" ],
    > [ tbls[1], tbls[2], isos[3], "2.L3(4).(2^2)_{123*}" ],
    > [ tbls[1], isos[2], isos[3], "2.L3(4).(2^2)_{12*3*}" ],
    > [ isos[1], tbls[2], tbls[3], "2.L3(4).(2^2)_{1*23}" ],
    > [ isos[1], isos[2], tbls[3], "2.L3(4).(2^2)_{1*2*3}" ],
    > [ isos[1], tbls[2], isos[3], "2.L3(4).(2^2)_{1*23*}" ],
    > [ isos[1], isos[2], isos[3], "2.L3(4).(2^2)_{1*2*3*}" ] ];;
    gap> tblG:= CharacterTable( "2.L3(4)" );;
    gap> result:= [];;
    gap> for input in inputs do
    >      tblsG2:= input{ [ 1 .. 3 ] };
    >      lib:= CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, input[4], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >      Append( result, RepresentativesCharacterTables( poss ) );
    >    od;
    #I  excluded cand. 2 (out of 8) for 2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 2.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 2.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 2.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 2.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 2.L3(4).(2^2)_{12*3*} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 2.L3(4).(2^2)_{12*3*} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 2.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 2.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 2.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 2.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 2.L3(4).(2^2)_{1*2*3} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 2.L3(4).(2^2)_{1*2*3} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 2.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 2.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 2.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 2.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 2.L3(4).(2^2)_{1*23*} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 2.L3(4).(2^2)_{1*23*} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 2.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 2.L3(4).(2^2)_{1*23*} by 5-mod. table
    gap> result:= List( result, x -> x.table );
    [ CharacterTable( "new2.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{12*3*}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{1*23}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{1*2*3}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{1*23*}" ), 
      CharacterTable( "new2.L3(4).(2^2)_{1*2*3*}" ) ]

We get exactly one character table for each input. For each of these tables, there are three possibilities to form an isoclinic table, corresponding to the three subgroups of index two. It turns out that the eight solutions form two orbits under forming some isoclinic table. Tables in different orbits are essentially different, already their numbers of conjugacy classes differ.
    gap> List( result, NrConjugacyClasses );
    [ 39, 33, 33, 39, 33, 39, 39, 33 ]
    gap> t:= result[1];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 4, 7, 6 ]
    gap> t:= result[2];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 3, 8, 5 ]

Up to now, it is not clear that the character tables we have constructed are really the character tables of some groups. The existence of groups for the first orbit of character tables can be established as follows.
The group U6(2).2 contains a maximal subgroup M of the type L3(4).22, see [CCN+85,p. 115]. Its derived subgroup M′ of the type L3(4) lies inside U6(2), and we claim that the preimage of M′ under the natural epimorphism from 2.U6(2) to U6(2) is a double cover of L3(4). Unfortunately, L3(4) admits class fusions into 2.U6(2), so this criterion cannot be used.
    gap> l34:= CharacterTable( "L3(4)" );;
    gap> u:= CharacterTable( "U6(2)" );;
    gap> 2u:= CharacterTable( "2.U6(2)" );;
    gap> cand:= PossibleClassFusions( l34, 2u );
    [ [ 1, 5, 12, 16, 22, 22, 23, 23, 41, 41 ], 
      [ 1, 5, 12, 22, 16, 22, 23, 23, 41, 41 ], 
      [ 1, 5, 12, 22, 22, 16, 23, 23, 41, 41 ] ]
    gap> OrdersClassRepresentatives( l34 );
    [ 1, 2, 3, 4, 4, 4, 5, 5, 7, 7 ]

Consider the three classes of elements of order four in L3(4). Under the possible fusions into 2.U6(2), they are mapped to the classes 16 and 22, which are preimages of the classes 10 and 14 (4C and 4G) of U6(2). Note that the maximal subgroups of type L3(4).2 in U6(2) extend to L3(4).6 type subgroups in U6(2).3, and the three classes 4C, 4D, 4E form one orbit under the action of an outer automorphism of order three of U6(2).
    gap> GetFusionMap( 2u, u ){ [ 16, 22 ] };
    [ 10, 14 ]
    gap> ClassNames( u, "ATLAS" ){ [ 10, 14 ] };
    [ "4C", "4G" ]
    gap> GetFusionMap( u, CharacterTable( "U6(2).3" ) );
    [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 10, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 
      20, 21, 22, 23, 24, 24, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 
      36, 36, 37, 38, 39, 40 ]

This means that any L3(4) type subgroup of U6(2) that extends to an L3(4).6 type subgroup in U6(2).3 either contains elements from all thrre classes 4C, 4D, 4E of U6(2), or contains no element from these classes. Thus we know that any double cover of U6(2).2 contains a double cover of L3(4).22. Only the first of our result tables admits a class fusion into the table of 2.U6(2).2.
    gap> 2u2:= CharacterTable( "2.U6(2).2" );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, 2u2 ) );;
    gap> List( fus, Length );
    [ 4, 0, 0, 0, 0, 0, 0, 0 ]

As a consequence, the fourth result table is established as that of a maximal subgroup of the group isoclinic to 2.U6(2).2.
    gap> 2u2iso:= CharacterTableIsoclinic( 2u2 );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, 2u2iso ) );;
    gap> List( fus, Length );
    [ 0, 0, 0, 4, 0, 0, 0, 0 ]

Similarly, the group HS.2 contains a maximal subgroup M of the type L3(4).22, see [CCN+85,p. 80]. Its derived subgroup M′ of the type L3(4) lies inside HS, and the preimage of M′ under the natural epimorphism from 2.HS to HS is a double cover of L3(4), because L3(4) does not admit a class fusion into 2.HS.2.
    gap> h2:= CharacterTable( "HS.2" );;
    gap> 2h2:= CharacterTable( "2.HS.2" );;
    gap> PossibleClassFusions( l34, 2h2 );
    [  ]

Only the fifth of our result tables admits a class fusion into 2.HS.2, which means that 2.L3(4).(22)1∗23 is a subgroup of 2.HS.2, and the eighth result table -2.L3(4).(22)1∗2∗3∗- admits a class fusion into the isoclinic variant of 2.HS.2 This shows the existence of groups for the tables from the second orbit.
    gap> fus:= List( result, x -> PossibleClassFusions( x, 2h2 ) );;
    gap> List( fus, Length );
    [ 0, 0, 0, 0, 4, 0, 0, 0 ]
    gap> 2h2iso:= CharacterTableIsoclinic( 2h2 );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, 2h2iso ) );;
    gap> List( fus, Length );
    [ 0, 0, 0, 0, 0, 0, 0, 4 ]

6.5  The Character Tables of Groups of the Type 6.L3(4).22 (October 2011)

We have two approaches for constructing the character tables of these groups.
First, we may regard them as normal products of the three normal subgroups of index two, each of them having the structure 6.L3(4).2, and use the approach from Section 3.3, as we did in Section 6.4 for the groups of the structure 2.L3(4).22.
Second, we may use the approach from Section 3.1. Note that the factor group L3(4).22 contains each of the three groups L3(4).2i as a subgroup, for 1 ≤ i ≤ 3, and the groups of the type 6.L3(4).21 have a centre of order six, whereas the centres of the 6.L3(4).22 and 6.L3(4).23 type groups have order two. For that, the character tables of the subgroups 6.L3(4).21 and 6.L3(4).21 are needed, as well as the character tables of the eight possible factor groups 2.L3(4).22; the latter tables are known from Section 6.4.
We show both approaches. (The second approach is better suited for storing the character tables in the Character Table Library, since the irreducible characters need not be stored, and since the Brauer tables of the groups can be derived from the Brauer tables of the compound tables.)
    gap> tbls:= List( [ "1", "2", "3" ],
    >      i -> CharacterTable( Concatenation( "6.L3(4).2_", i ) ) );
    [ CharacterTable( "6.L3(4).2_1" ), CharacterTable( "6.L3(4).2_2" ), 
      CharacterTable( "6.L3(4).2_3" ) ]
    gap> isos:= List( [ "1", "2", "3" ],
    >      i -> CharacterTable( Concatenation( "6.L3(4).2_", i, "*" ) ) );
    [ CharacterTable( "Isoclinic(6.L3(4).2_1)" ), 
      CharacterTable( "Isoclinic(6.L3(4).2_2)" ), 
      CharacterTable( "Isoclinic(6.L3(4).2_3)" ) ]
    gap> inputs:= [
    > [ tbls[1], tbls[2], tbls[3], "6.L3(4).(2^2)_{123}" ],
    > [ tbls[1], isos[2], tbls[3], "6.L3(4).(2^2)_{12*3}" ],
    > [ tbls[1], tbls[2], isos[3], "6.L3(4).(2^2)_{123*}" ],
    > [ tbls[1], isos[2], isos[3], "6.L3(4).(2^2)_{12*3*}" ],
    > [ isos[1], tbls[2], tbls[3], "6.L3(4).(2^2)_{1*23}" ],
    > [ isos[1], isos[2], tbls[3], "6.L3(4).(2^2)_{1*2*3}" ],
    > [ isos[1], tbls[2], isos[3], "6.L3(4).(2^2)_{1*23*}" ],
    > [ isos[1], isos[2], isos[3], "6.L3(4).(2^2)_{1*2*3*}" ] ];;
    gap> tblG:= CharacterTable( "6.L3(4)" );;
    gap> result:= [];;
    gap> for input in inputs do
    >      tblsG2:= input{ [ 1 .. 3 ] };
    >      lib:= CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, input[4], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >      Append( result, RepresentativesCharacterTables( poss ) );
    >    od;
    #I  excluded cand. 2 (out of 8) for 6.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 6.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 6.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 6.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 6.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 6.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 6.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 6.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 6.L3(4).(2^2)_{12*3*} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 6.L3(4).(2^2)_{12*3*} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 6.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 6.L3(4).(2^2)_{12*3*} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 6.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 6.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 6.L3(4).(2^2)_{1*2*3} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 6.L3(4).(2^2)_{1*2*3} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 6.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 6.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 6.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 6.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 6.L3(4).(2^2)_{1*23*} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 6.L3(4).(2^2)_{1*23*} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 6.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 6.L3(4).(2^2)_{1*23*} by 5-mod. table
    gap> result:= List( result, x -> x.table );
    [ CharacterTable( "new6.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{12*3*}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*23}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*2*3}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*23*}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*2*3*}" ) ]

As in Section 6.4, we get exactly one character table for each input, and the eight solutions lie in two orbits under isoclinism.
    gap> List( result, NrConjugacyClasses );
    [ 59, 53, 53, 59, 53, 59, 59, 53 ]
    gap> t:= result[1];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 7, 6, 4 ]
    gap> t:= result[2];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 8, 5, 3 ]

Up to now, it is not clear that the character tables we have constructed are really the character tables of some groups. The existence of groups for the first orbit of character tables can be established as follows.
We have shown in Section 6.4 that the maximal subgroups M of the type L3(4).22 in U6(2).2 lift to double covers 2.L3(4).22 in 2.U6(2).2. The preimages of these groups under the natural epimorphism from 6.U6(2).2 have the structure 6.L3(4).22, where the derived subgroup is the six-fold cover of L3(4); this follows from the fact that 6.U6(2) does not admit a class fusion from the double cover 2.L3(4).
    gap> 2l34:= CharacterTable( "2.L3(4)" );;
    gap> 6u:= CharacterTable( "6.U6(2)" );;
    gap> cand:= PossibleClassFusions( 2l34, 6u );
    [  ]

This establishes the first and the fourth result as character tables of subgroups of 6.U6(2) and its isoclinic variant, respectively.
    gap> 6u2:= CharacterTable( "6.U6(2).2" );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, 6u2 ) );;
    gap> List( fus, Length );
    [ 8, 0, 0, 0, 0, 0, 0, 0 ]
    gap> 6u2iso:= CharacterTableIsoclinic( 6u2 );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, 6u2iso ) );;
    gap> List( fus, Length );
    [ 0, 0, 0, 8, 0, 0, 0, 0 ]

Similarly, the group G2(4).2 contains a maximal subgroup M of the type 3.L3(4).22, see [CCN+85,p. 97]. Its derived subgroup M′ of the type 3.L3(4) lies inside G2(4), and the preimage of M′ under the natural epimorphism from 2.G2(4) to G2(4) is a double cover of 3.L3(4), because 3.L3(4) does not admit a class fusion into 2.G2(4).2.
    gap> 3l34:= CharacterTable( "3.L3(4)" );;
    gap> g2:= CharacterTable( "G2(4).2" );;
    gap> 2g2:= CharacterTable( "2.G2(4).2" );;
    gap> PossibleClassFusions( 3l34, 2g2 );
    [  ]

Only the third and eighth of our result tables admit a class fusion into 2.G2(4).2 and its isoclinic variant, respectively. This shows the existence of groups for the tables from the second orbit.
    gap> fus:= List( result, x -> PossibleClassFusions( x, 2g2 ) );;
    gap> List( fus, Length );
    [ 0, 0, 16, 0, 0, 0, 0, 0 ]
    gap> 2g2iso:= CharacterTableIsoclinic( 2g2 );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, 2g2iso ) );;
    gap> List( fus, Length );
    [ 0, 0, 0, 0, 0, 0, 0, 16 ]

Now we try the second approach and compare the results.
    gap> names:= [ "L3(4).(2^2)_{123}", "L3(4).(2^2)_{12*3}",
    >              "L3(4).(2^2)_{123*}", "L3(4).(2^2)_{12*3*}" ];;
    gap> inputs1:= List( names, nam -> [ "6.L3(4).2_1", "2.L3(4).2_1",
    >                Concatenation( "2.", nam ), Concatenation( "6.", nam ) ] );;
    gap> names:= List( names, nam -> ReplacedString( nam, "1", "1*" ) );;
    gap> inputs2:= List( names, nam -> [ "6.L3(4).2_1*", "2.L3(4).2_1*",
    >                Concatenation( "2.", nam ), Concatenation( "6.", nam ) ] );;
    gap> inputs:= Concatenation( inputs1, inputs2 );
    [ [ "6.L3(4).2_1", "2.L3(4).2_1", "2.L3(4).(2^2)_{123}", 
          "6.L3(4).(2^2)_{123}" ], 
      [ "6.L3(4).2_1", "2.L3(4).2_1", "2.L3(4).(2^2)_{12*3}", 
          "6.L3(4).(2^2)_{12*3}" ], 
      [ "6.L3(4).2_1", "2.L3(4).2_1", "2.L3(4).(2^2)_{123*}", 
          "6.L3(4).(2^2)_{123*}" ], 
      [ "6.L3(4).2_1", "2.L3(4).2_1", "2.L3(4).(2^2)_{12*3*}", 
          "6.L3(4).(2^2)_{12*3*}" ], 
      [ "6.L3(4).2_1*", "2.L3(4).2_1*", "2.L3(4).(2^2)_{1*23}", 
          "6.L3(4).(2^2)_{1*23}" ], 
      [ "6.L3(4).2_1*", "2.L3(4).2_1*", "2.L3(4).(2^2)_{1*2*3}", 
          "6.L3(4).(2^2)_{1*2*3}" ], 
      [ "6.L3(4).2_1*", "2.L3(4).2_1*", "2.L3(4).(2^2)_{1*23*}", 
          "6.L3(4).(2^2)_{1*23*}" ], 
      [ "6.L3(4).2_1*", "2.L3(4).2_1*", "2.L3(4).(2^2)_{1*2*3*}", 
          "6.L3(4).(2^2)_{1*2*3*}" ] ]
    gap> result2:= [];;
    gap> for  input in inputs do
    >      tblMG := CharacterTable( input[1] );
    >      tblG  := CharacterTable( input[2] );
    >      tblGA := CharacterTable( input[3] );
    >      name  := Concatenation( "new", input[4] );
    >      lib   := CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );
    >      Append( result2, poss );
    >    od;
    gap> result2:= List( result2, x -> x.table );
    [ CharacterTable( "new6.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{12*3*}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*23}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*2*3}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*23*}" ), 
      CharacterTable( "new6.L3(4).(2^2)_{1*2*3*}" ) ]
    gap> trans:= List( [ 1 .. 8 ], i ->
    >        TransformingPermutationsCharacterTables( result[i], result2[i] ) );;
    gap> ForAll( trans, IsRecord );
    true

6.6  The Character Tables of Groups of the Type 2.U4(3).22 (February 2012)

The outer automorphism group of the group U4(3) is a dihedral group of order 8. There are two almost simple groups of the type U4(3).22, up to isomorphism, denoted as U4(3).(22)122 and U4(3).(22)133, respectively. Note that U4(3).21 is the extension by the central involution of the outer automorphism group of U4(3), the other two subgroups of index two in U4(3).(22)122 are U4(3).22 and U4(3).22, respectively, and the other two subgroups of index two in U4(3).(22)133 are U4(3).23 and U4(3).23, respectively.
Since Aut( U4(3) ) possesses a double cover (see [CCN+85,p. 52]), double covers of U4(3).(22)122 and U4(3).(22)133 exist.
First we deal with the double covers of U4(3).(22)122. Any such group contains one subgroup of the type 2.U4(3).21 and two subgroups of the type 2.U4(3).22, and there are two isoclinic variants of each of these group to consider, see Section 2.5. Thus we start with six different inputs for the construction of the character tables of double covers.
    gap> tbls:= List( [ "1", "2", "2'" ], i ->
    >      CharacterTable( Concatenation( "2.U4(3).2_", i ) ) );;
    gap> isos:= List( [ "1", "2", "2'" ], i ->
    >      CharacterTable( Concatenation( "Isoclinic(2.U4(3).2_", i, ")" ) ) );;
    gap> inputs:= [
    > [ tbls[1], tbls[2], tbls[3], "2.U4(3).(2^2)_{122}" ],
    > [ isos[1], tbls[2], tbls[3], "2.U4(3).(2^2)_{1*22}" ],
    > [ tbls[1], isos[2], tbls[3], "2.U4(3).(2^2)_{12*2}" ],
    > [ isos[1], isos[2], tbls[3], "2.U4(3).(2^2)_{1*2*2}" ],
    > [ tbls[1], isos[2], isos[3], "2.U4(3).(2^2)_{12*2*}" ],
    > [ isos[1], isos[2], isos[3], "2.U4(3).(2^2)_{1*2*2*}" ] ];;
    gap> tblG:= CharacterTable( "2.U4(3)" );;
    gap> result:= [];;
    gap> for input in inputs do
    >      tblsG2:= input{ [ 1 .. 3 ] };
    >      lib:= CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, input[4], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >      Append( result, RepresentativesCharacterTables( poss ) );
    >    od;
    gap> result:= List( result, x -> x.table );
    [ CharacterTable( "new2.U4(3).(2^2)_{122}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{1*22}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{12*2}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{1*2*2}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{12*2*}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{1*2*2*}" ) ]

We get exactly one character table for each input. For each of these tables, there are three possibilities to form an isoclinic table, corresponding to the three subgroups of index two. It turns out that the six solutions form two orbits under forming some isoclinic table. Tables in different orbits are essentially different, already their numbers of conjugacy classes differ.
    gap> List( result, NrConjugacyClasses );
    [ 87, 102, 102, 87, 87, 102 ]
    gap> t:= result[1];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 4, 4, 5 ]
    gap> t:= result[2];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 3, 3, 6 ]

Up to now, it is not clear that the character tables we have constructed are really the character tables of some groups. The existence of groups for the first orbit of character tables can be established as follows.
The group O8+(3) contains maximal subgroups of the type 2.U4(3).22, see [CCN+85,p. 140]. Only the first of our result tables admits a class fusion into the table of O8+(3).
    gap> u:= CharacterTable( "O8+(3)" );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, u ) );;
    gap> List( fus, Length );
    [ 24, 0, 0, 0, 0, 0 ]

A table in the second orbit belongs to a maximal subgroup of O7(3).2, see [CCN+85,p. 109].
    gap> u:= CharacterTable( "O7(3).2" );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, u ) );;
    gap> List( fus, Length );
    [ 0, 16, 0, 0, 0, 0 ]

Note that this subgroup of O7(3).2 ≅ SO(7,3) is the orthogonal group GO6(3).
Now we deal with the double covers of U4(3).(22)133. The constructions of the character tables are completely analogous to those in the case of U4(3).(22)122.
    gap> tbls:= List( [ "1", "3", "3'" ],
    >      i -> CharacterTable( Concatenation( "2.U4(3).2_", i ) ) );;
    gap> isos:= List( [ "1", "3", "3'" ], i ->
    >      CharacterTable( Concatenation( "Isoclinic(2.U4(3).2_", i, ")" ) ) );;
    gap> inputs:= [
    > [ tbls[1], tbls[2], tbls[3], "2.U4(3).(2^2)_{133}" ],
    > [ isos[1], tbls[2], tbls[3], "2.U4(3).(2^2)_{1*33}" ],
    > [ tbls[1], isos[2], tbls[3], "2.U4(3).(2^2)_{13*3}" ],
    > [ isos[1], isos[2], tbls[3], "2.U4(3).(2^2)_{1*3*3}" ],
    > [ tbls[1], isos[2], isos[3], "2.U4(3).(2^2)_{13*3*}" ],
    > [ isos[1], isos[2], isos[3], "2.U4(3).(2^2)_{1*3*3*}" ] ];;
    gap> tblG:= CharacterTable( "2.U4(3)" );;
    gap> result:= [];;
    gap> for input in inputs do
    >      tblsG2:= input{ [ 1 .. 3 ] };
    >      lib:= CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, input[4], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >      Append( result, RepresentativesCharacterTables( poss ) );
    >    od;
    #I  excluded cand. 2 (out of 4) for 2.U4(3).(2^2)_{1*33} by 3-mod. table
    #I  excluded cand. 3 (out of 4) for 2.U4(3).(2^2)_{1*33} by 3-mod. table
    #I  excluded cand. 2 (out of 4) for 2.U4(3).(2^2)_{13*3} by 3-mod. table
    #I  excluded cand. 3 (out of 4) for 2.U4(3).(2^2)_{13*3} by 3-mod. table
    #I  excluded cand. 2 (out of 4) for 2.U4(3).(2^2)_{1*3*3*} by 3-mod. table
    #I  excluded cand. 3 (out of 4) for 2.U4(3).(2^2)_{1*3*3*} by 3-mod. table
    gap> result:= List( result, x -> x.table );
    [ CharacterTable( "new2.U4(3).(2^2)_{133}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{1*33}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{13*3}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{1*3*3}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{13*3*}" ), 
      CharacterTable( "new2.U4(3).(2^2)_{1*3*3*}" ) ]
    gap> List( result, NrConjugacyClasses );
    [ 69, 72, 72, 69, 69, 72 ]
    gap> t:= result[1];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 4, 4, 5 ]
    gap> t:= result[2];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 3, 3, 6 ]

6.7  The Character Tables of Groups of the Type 41.L3(4).22 (October 2011)

The situation with 41.L3(4).22 is analogous to that with 6.L3(4).22, see Section 6.5.
    gap> tbls:= List( [ "1", "2", "3" ],
    >      i -> CharacterTable( Concatenation( "4_1.L3(4).2_", i ) ) );
    [ CharacterTable( "4_1.L3(4).2_1" ), CharacterTable( "4_1.L3(4).2_2" ), 
      CharacterTable( "4_1.L3(4).2_3" ) ]
    gap> isos:= List( [ "1", "2", "3" ],
    >      i -> CharacterTable( Concatenation( "4_1.L3(4).2_", i, "*" ) ) );
    [ CharacterTable( "Isoclinic(4_1.L3(4).2_1)" ), 
      CharacterTable( "Isoclinic(4_1.L3(4).2_2)" ), 
      CharacterTable( "4_1.L3(4).2_3*" ) ]

Note that the group 41.L3(4).23 has a centre of order four, so one cannot construct the isoclinic variant by calling the one argument version of CharacterTableIsoclinic.
    gap> List( tbls, ClassPositionsOfCentre );
    [ [ 1, 3 ], [ 1, 3 ], [ 1, 2, 3, 4 ] ]
    gap> IsRecord( TransformingPermutationsCharacterTables( tbls[3],
    >        CharacterTableIsoclinic( tbls[3] ) ) );
    true

Again, we get eight different character tables, in two orbits.
    gap> inputs:= [
    > [ tbls[1], tbls[2], tbls[3], "4_1.L3(4).(2^2)_{123}" ],
    > [ isos[1], tbls[2], tbls[3], "4_1.L3(4).(2^2)_{1*23}" ],
    > [ tbls[1], isos[2], tbls[3], "4_1.L3(4).(2^2)_{12*3}" ],
    > [ isos[1], isos[2], tbls[3], "4_1.L3(4).(2^2)_{1*2*3}" ],
    > [ tbls[1], tbls[2], isos[3], "4_1.L3(4).(2^2)_{123*}" ],
    > [ isos[1], tbls[2], isos[3], "4_1.L3(4).(2^2)_{1*23*}" ],
    > [ tbls[1], isos[2], isos[3], "4_1.L3(4).(2^2)_{12*3*}" ],
    > [ isos[1], isos[2], isos[3], "4_1.L3(4).(2^2)_{1*2*3*}" ] ];;
    gap> tblG:= CharacterTable( "4_1.L3(4)" );;
    gap> result:= [];;
    gap> for input in inputs do
    >      tblsG2:= input{ [ 1 .. 3 ] };
    >      lib:= CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, input[4], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >      Append( result, RepresentativesCharacterTables( poss ) );
    >    od;
    #I  excluded cand. 2 (out of 8) for 4_1.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_1.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_1.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_1.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_1.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_1.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 4_1.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_1.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_1.L3(4).(2^2)_{1*23} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_1.L3(4).(2^2)_{1*23} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_1.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_1.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 4_1.L3(4).(2^2)_{12*3} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_1.L3(4).(2^2)_{12*3} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_1.L3(4).(2^2)_{12*3} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_1.L3(4).(2^2)_{12*3} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_1.L3(4).(2^2)_{12*3} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_1.L3(4).(2^2)_{12*3} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 4_1.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_1.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_1.L3(4).(2^2)_{1*2*3} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_1.L3(4).(2^2)_{1*2*3} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_1.L3(4).(2^2)_{1*2*3} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_1.L3(4).(2^2)_{1*2*3} by 5-mod. table
    gap> result:= List( result, x -> x.table );
    [ CharacterTable( "new4_1.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{1*23}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{1*2*3}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{1*23*}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{12*3*}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{1*2*3*}" ) ]
    gap> List( result, NrConjugacyClasses );
    [ 48, 48, 48, 48, 42, 42, 42, 42 ]
    gap> t:= result[1];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 3, 2, 4 ]
    gap> t:= result[5];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 7, 6, 8 ]

Note that only two out of the eight tables of the type 2.L3(4).22 occur as factors of the eight tables.
    gap> facts:= [ CharacterTable( "2.L3(4).(2^2)_{123}" ), 
    >              CharacterTable( "2.L3(4).(2^2)_{123*}" ) ];;
    gap> factresults:= List( result, t -> t / ClassPositionsOfCentre( t ) );;
    gap> List( factresults, t -> PositionProperty( facts, f ->
    >            IsRecord( TransformingPermutationsCharacterTables( t, f ) ) ) );
    [ 1, 1, 1, 1, 2, 2, 2, 2 ]

This is not surprising; note that for 1 ≤ i ≤ 2, the two isoclinic variants of 41.L3(4).2i have isomorphic factor groups of the type 2.L3(4).2i. (For i = 3, this is not the case.)
    gap> test:= [ CharacterTable( "4_1.L3(4).2_1" ),
    >             CharacterTable( "4_1.L3(4).2_1*" ) ];;
    gap> List( test, ClassPositionsOfCentre );
    [ [ 1, 3 ], [ 1, 3 ] ]
    gap> fact:= List( test, t -> t / ClassPositionsOfCentre( t ) );;
    gap> IsRecord( TransformingPermutationsCharacterTables( fact[1], fact[2] ) );
    true
    gap> test:= [ CharacterTable( "4_1.L3(4).2_2" ),
    >             CharacterTable( "4_1.L3(4).2_2*" ) ];;
    gap> List( test, ClassPositionsOfCentre );
    [ [ 1, 3 ], [ 1, 3 ] ]
    gap> fact:= List( test, t -> t / ClassPositionsOfCentre( t ) );;
    gap> IsRecord( TransformingPermutationsCharacterTables( fact[1], fact[2] ) );
    true

Now we try the second approach and compare the results. By the abovementioned asymmetry, it is clear that the tables are not uniquely determined by the input data.
    gap> names:= [ "L3(4).(2^2)_{123}", "L3(4).(2^2)_{1*23}",
    >              "L3(4).(2^2)_{12*3}", "L3(4).(2^2)_{1*2*3}" ];;
    gap> inputs1:= List( names, nam -> [ "4_1.L3(4).2_3", "2.L3(4).2_3",
    >      Concatenation( "2.", nam ), Concatenation( "4_1.", nam ) ] );;
    gap> names:= List( names, nam -> ReplacedString( nam, "3}", "3*}" ) );;
    gap> inputs2:= List( names, nam -> [ "4_1.L3(4).2_3*", "2.L3(4).2_3*",
    >      Concatenation( "2.", nam ), Concatenation( "4_1.", nam ) ] );;
    gap> inputs:= Concatenation( inputs1, inputs2 );
    [ [ "4_1.L3(4).2_3", "2.L3(4).2_3", "2.L3(4).(2^2)_{123}", 
          "4_1.L3(4).(2^2)_{123}" ], 
      [ "4_1.L3(4).2_3", "2.L3(4).2_3", "2.L3(4).(2^2)_{1*23}", 
          "4_1.L3(4).(2^2)_{1*23}" ], 
      [ "4_1.L3(4).2_3", "2.L3(4).2_3", "2.L3(4).(2^2)_{12*3}", 
          "4_1.L3(4).(2^2)_{12*3}" ], 
      [ "4_1.L3(4).2_3", "2.L3(4).2_3", "2.L3(4).(2^2)_{1*2*3}", 
          "4_1.L3(4).(2^2)_{1*2*3}" ], 
      [ "4_1.L3(4).2_3*", "2.L3(4).2_3*", "2.L3(4).(2^2)_{123*}", 
          "4_1.L3(4).(2^2)_{123*}" ], 
      [ "4_1.L3(4).2_3*", "2.L3(4).2_3*", "2.L3(4).(2^2)_{1*23*}", 
          "4_1.L3(4).(2^2)_{1*23*}" ], 
      [ "4_1.L3(4).2_3*", "2.L3(4).2_3*", "2.L3(4).(2^2)_{12*3*}", 
          "4_1.L3(4).(2^2)_{12*3*}" ], 
      [ "4_1.L3(4).2_3*", "2.L3(4).2_3*", "2.L3(4).(2^2)_{1*2*3*}", 
          "4_1.L3(4).(2^2)_{1*2*3*}" ] ]
    gap> result2:= [];;
    gap> for  input in inputs do
    >      tblMG := CharacterTable( input[1] );
    >      tblG  := CharacterTable( input[2] );
    >      tblGA := CharacterTable( input[3] );
    >      name  := Concatenation( "new", input[4] );
    >      lib   := CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );
    >      Append( result2, poss );
    >    od;
    #E  4 possibilities for new4_1.L3(4).(2^2)_{123}
    #E  no solution for new4_1.L3(4).(2^2)_{1*23}
    #E  no solution for new4_1.L3(4).(2^2)_{12*3}
    #E  no solution for new4_1.L3(4).(2^2)_{1*2*3}
    #E  4 possibilities for new4_1.L3(4).(2^2)_{123*}
    #E  no solution for new4_1.L3(4).(2^2)_{1*23*}
    #E  no solution for new4_1.L3(4).(2^2)_{12*3*}
    #E  no solution for new4_1.L3(4).(2^2)_{1*2*3*}
    gap> Length( result2 );
    8
    gap> result2:= List( result2, x -> x.table );
    [ CharacterTable( "new4_1.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new4_1.L3(4).(2^2)_{123*}" ) ]
    gap> List( result, t1 -> PositionsProperty( result2, t2 -> IsRecord(
    >      TransformingPermutationsCharacterTables( t1, t2 ) ) ) );
    [ [ 1 ], [ 4 ], [ 3 ], [ 2 ], [ 7 ], [ 6 ], [ 5 ], [ 8 ] ]

At the moment, I do not know interesting groups that contain one of the 41.L3(4).22 type groups and whose character tables are available.

6.8  The Character Tables of Groups of the Type 42.L3(4).22 (October 2011)

The situation with 42.L3(4).22 is analogous to that with 6.L3(4).22, see Section 6.5.
    gap> tbls:= List( [ "1", "2", "3" ],
    >      i -> CharacterTable( Concatenation( "4_2.L3(4).2_", i ) ) );
    [ CharacterTable( "4_2.L3(4).2_1" ), CharacterTable( "4_2.L3(4).2_2" ), 
      CharacterTable( "4_2.L3(4).2_3" ) ]
    gap> isos:= List( [ "1", "2", "3" ], 
    >      i -> CharacterTable( Concatenation( "4_2.L3(4).2_", i, "*" ) ) );
    [ CharacterTable( "Isoclinic(4_2.L3(4).2_1)" ), 
      CharacterTable( "4_2.L3(4).2_2*" ), 
      CharacterTable( "Isoclinic(4_2.L3(4).2_3)" ) ]

Note that the group 41.L3(4).22 has a centre of order four, so one cannot construct the isoclinic variant not by calling the one argument version of CharacterTableIsoclinic.
    gap> List( tbls, ClassPositionsOfCentre );
    [ [ 1, 3 ], [ 1, 2, 3, 4 ], [ 1, 3 ] ]
    gap> IsRecord( TransformingPermutationsCharacterTables( tbls[2],
    >        CharacterTableIsoclinic( tbls[2] ) ) );
    true

Again, we get eight different character tables, in two orbits.
    gap> inputs:= [
    > [ tbls[1], tbls[2], tbls[3], "4_2.L3(4).(2^2)_{123}" ],
    > [ isos[1], tbls[2], tbls[3], "4_2.L3(4).(2^2)_{1*23}" ],
    > [ tbls[1], isos[2], tbls[3], "4_2.L3(4).(2^2)_{12*3}" ],
    > [ tbls[1], tbls[2], isos[3], "4_2.L3(4).(2^2)_{123*}" ],
    > [ isos[1], isos[2], tbls[3], "4_2.L3(4).(2^2)_{1*2*3}" ],
    > [ isos[1], tbls[2], isos[3], "4_2.L3(4).(2^2)_{1*23*}" ],
    > [ tbls[1], isos[2], isos[3], "4_2.L3(4).(2^2)_{12*3*}" ],
    > [ isos[1], isos[2], isos[3], "4_2.L3(4).(2^2)_{1*2*3*}" ] ];;
    gap> tblG:= CharacterTable( "4_2.L3(4)" );;
    gap> result:= [];;
    gap> for input in inputs do
    >      tblsG2:= input{ [ 1 .. 3 ] };
    >      lib:= CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryGV4Table( tblG, tblsG2, input[4], lib );
    >      ConstructModularGV4Tables( tblG, tblsG2, poss, lib );
    >      Append( result, RepresentativesCharacterTables( poss ) );
    >    od;
    #I  excluded cand. 2 (out of 8) for 4_2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_2.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_2.L3(4).(2^2)_{123} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_2.L3(4).(2^2)_{123} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 4_2.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_2.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_2.L3(4).(2^2)_{1*23} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_2.L3(4).(2^2)_{1*23} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_2.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_2.L3(4).(2^2)_{1*23} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 4_2.L3(4).(2^2)_{123*} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_2.L3(4).(2^2)_{123*} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_2.L3(4).(2^2)_{123*} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_2.L3(4).(2^2)_{123*} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_2.L3(4).(2^2)_{123*} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_2.L3(4).(2^2)_{123*} by 5-mod. table
    #I  excluded cand. 2 (out of 8) for 4_2.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 3 (out of 8) for 4_2.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 4 (out of 8) for 4_2.L3(4).(2^2)_{1*23*} by 7-mod. table
    #I  excluded cand. 5 (out of 8) for 4_2.L3(4).(2^2)_{1*23*} by 7-mod. table
    #I  excluded cand. 6 (out of 8) for 4_2.L3(4).(2^2)_{1*23*} by 5-mod. table
    #I  excluded cand. 7 (out of 8) for 4_2.L3(4).(2^2)_{1*23*} by 5-mod. table
    gap> result:= List( result, x -> x.table );
    [ CharacterTable( "new4_2.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{1*23}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{123*}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{1*2*3}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{1*23*}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{12*3*}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{1*2*3*}" ) ]
    gap> List( result, NrConjugacyClasses );
    [ 50, 50, 44, 50, 44, 50, 44, 44 ]
    gap> t:= result[1];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 4, 2, 6 ]
    gap> t:= result[3];;
    gap> nsg:= Filtered( ClassPositionsOfNormalSubgroups( t ),
    >            x -> Sum( SizesConjugacyClasses( t ){ x } ) = Size( t ) / 2 );;
    gap> iso:= List( nsg, x -> CharacterTableIsoclinic( t, x ) );;
    gap> List( iso, x -> PositionProperty( result, y ->
    >            TransformingPermutationsCharacterTables( x, y ) <> fail ) );
    [ 7, 5, 8 ]

Note that only two out of the eight tables of the type 2.L3(4).22 occur as factors of the eight tables.
    gap> facts:= [ CharacterTable( "2.L3(4).(2^2)_{123}" ),
    >              CharacterTable( "2.L3(4).(2^2)_{12*3}" ) ];;
    gap> factresults:= List( result, t -> t / ClassPositionsOfCentre( t ) );;
    gap> List( factresults, t -> PositionProperty( facts, f ->
    >            IsRecord( TransformingPermutationsCharacterTables( t, f ) ) ) );
    [ 1, 1, 2, 1, 2, 1, 2, 2 ]

This is not surprising; note that for i ∈ { 1, 3 }, the two isoclinic variants of 41.L3(4).2i have isomorphic factor groups of the type 2.L3(4).2i. (For i = 2, this is not the case.)
    gap> test:= [ CharacterTable( "4_2.L3(4).2_1" ),
    >             CharacterTable( "4_2.L3(4).2_1*" ) ];;
    gap> List( test, ClassPositionsOfCentre );
    [ [ 1, 3 ], [ 1, 3 ] ]
    gap> fact:= List( test, t -> t / ClassPositionsOfCentre( t ) );;
    gap> IsRecord( TransformingPermutationsCharacterTables( fact[1], fact[2] ) );
    true
    gap> test:= [ CharacterTable( "4_2.L3(4).2_3" ),
    >             CharacterTable( "4_2.L3(4).2_3*" ) ];;
    gap> List( test, ClassPositionsOfCentre );
    [ [ 1, 3 ], [ 1, 3 ] ]
    gap> fact:= List( test, t -> t / ClassPositionsOfCentre( t ) );;
    gap> IsRecord( TransformingPermutationsCharacterTables( fact[1], fact[2] ) );
    true

Now we try the second approach and compare the results. By the abovementioned asymmetry, it is clear that the tables are not uniquely determined by the input data.
    gap> names:= [ "L3(4).(2^2)_{123}", "L3(4).(2^2)_{1*23}",
    >              "L3(4).(2^2)_{123*}", "L3(4).(2^2)_{1*23*}" ];;
    gap> inputs1:= List( names, nam -> [ "4_2.L3(4).2_2", "2.L3(4).2_2",
    >      Concatenation( "2.", nam ), Concatenation( "4_2.", nam ) ] );;
    gap> names:= List( names, nam -> ReplacedString( nam, "23", "2*3" ) );;
    gap> inputs2:= List( names, nam -> [ "4_2.L3(4).2_2*", "2.L3(4).2_2*",
    >      Concatenation( "2.", nam ), Concatenation( "4_2.", nam ) ] );;
    gap> inputs:= Concatenation( inputs1, inputs2 );
    [ [ "4_2.L3(4).2_2", "2.L3(4).2_2", "2.L3(4).(2^2)_{123}", 
          "4_2.L3(4).(2^2)_{123}" ], 
      [ "4_2.L3(4).2_2", "2.L3(4).2_2", "2.L3(4).(2^2)_{1*23}", 
          "4_2.L3(4).(2^2)_{1*23}" ], 
      [ "4_2.L3(4).2_2", "2.L3(4).2_2", "2.L3(4).(2^2)_{123*}", 
          "4_2.L3(4).(2^2)_{123*}" ], 
      [ "4_2.L3(4).2_2", "2.L3(4).2_2", "2.L3(4).(2^2)_{1*23*}", 
          "4_2.L3(4).(2^2)_{1*23*}" ], 
      [ "4_2.L3(4).2_2*", "2.L3(4).2_2*", "2.L3(4).(2^2)_{12*3}", 
          "4_2.L3(4).(2^2)_{12*3}" ], 
      [ "4_2.L3(4).2_2*", "2.L3(4).2_2*", "2.L3(4).(2^2)_{1*2*3}", 
          "4_2.L3(4).(2^2)_{1*2*3}" ], 
      [ "4_2.L3(4).2_2*", "2.L3(4).2_2*", "2.L3(4).(2^2)_{12*3*}", 
          "4_2.L3(4).(2^2)_{12*3*}" ], 
      [ "4_2.L3(4).2_2*", "2.L3(4).2_2*", "2.L3(4).(2^2)_{1*2*3*}", 
          "4_2.L3(4).(2^2)_{1*2*3*}" ] ]
    gap> result2:= [];;
    gap> for  input in inputs do
    >      tblMG := CharacterTable( input[1] );
    >      tblG  := CharacterTable( input[2] );
    >      tblGA := CharacterTable( input[3] );
    >      name  := Concatenation( "new", input[4] );
    >      lib   := CharacterTable( input[4] );
    >      poss:= ConstructOrdinaryMGATable( tblMG, tblG, tblGA, name, lib );
    >      Append( result2, poss );
    >    od;
    #E  4 possibilities for new4_2.L3(4).(2^2)_{123}
    #E  no solution for new4_2.L3(4).(2^2)_{1*23}
    #E  no solution for new4_2.L3(4).(2^2)_{123*}
    #E  no solution for new4_2.L3(4).(2^2)_{1*23*}
    #E  4 possibilities for new4_2.L3(4).(2^2)_{12*3}
    #E  no solution for new4_2.L3(4).(2^2)_{1*2*3}
    #E  no solution for new4_2.L3(4).(2^2)_{12*3*}
    #E  no solution for new4_2.L3(4).(2^2)_{1*2*3*}
    gap> Length( result2 );
    8
    gap> result2:= List( result2, x -> x.table );
    [ CharacterTable( "new4_2.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{123}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{12*3}" ), 
      CharacterTable( "new4_2.L3(4).(2^2)_{12*3}" ) ]
    gap> List( result, t1 -> PositionsProperty( result2, t2 -> IsRecord(
    >      TransformingPermutationsCharacterTables( t1, t2 ) ) ) );
    [ [ 1 ], [ 4 ], [ 7 ], [ 3 ], [ 6 ], [ 2 ], [ 5 ], [ 8 ] ]

The group ON.2 contains a maximal subgroup M of the type 42.L3(4).22, see [CCN+85,p. 132]. Only the third result table admits a class fusion into ON.2. This shows the existence of groups for the tables from the second orbit.
    gap> on2:= CharacterTable( "ON.2" );;
    gap> fus:= List( result, x -> PossibleClassFusions( x, on2 ) );;
    gap> List( fus, Length );
    [ 0, 0, 16, 0, 0, 0, 0, 0 ]

6.9  The Character Table of Aut(L2(81))

The group Aut(L2(81)) ≅ L2(81).(2 ×4) has the structure G.22 where G = L2(81).21. Here we get two triples of possible actions on the tables of the groups G.2i, and one possible character table for each triple.
    gap> input:= [ "L2(81).2_1", "L2(81).4_1", "L2(81).4_2", "L2(81).2^2",
    >                                                        "L2(81).(2x4)" ];;
    gap> tblG   := CharacterTable( input[1] );;
    gap> tblsG2 := List( input{ [ 2 .. 4 ] }, CharacterTable );;
    gap> name   := Concatenation( "new", input[5] );;
    gap> lib    := CharacterTable( input[5] );;
    gap> poss   := ConstructOrdinaryGV4Table( tblG, tblsG2, name, lib );;
    #I  newL2(81).(2x4): 2 equivalence classes
    gap> reps:= RepresentativesCharacterTables( poss );;
    gap> Length( reps );
    2

Due to the different underlying actions, the power maps of the two candidate tables differ.
    gap> ord:= OrdersClassRepresentatives( reps[1].table );;
    gap> ord = OrdersClassRepresentatives( reps[2].table ); 
    true
    gap> pos:= Position( ord, 80 );
    33
    gap> PowerMap( reps[1].table, 3 )[ pos ];
    34
    gap> PowerMap( reps[2].table, 3 )[ pos ];
    33

Aut(L2(81)) can be generated by PGL(2,81) = L2(81).22 and the Frobenius automorphism of order four that is defined on GL(2,81) as the map that cubes the matrix entries. The elements of order 80 in Aut(L2(81)) are conjugates of diagonal matrices modulo scalar matrices, which are mapped to their third powers by the Frobenius homomorphism. So the third power map of Aut(L2(81)) fixes the classes of elements of order 80. In other words, the second of the two tables is the right one.
    gap> trans:= TransformingPermutationsCharacterTables( reps[2].table, lib );;
    gap> IsRecord( trans );
    true
    gap> List( reps[2].G2fusGV4, x -> OnTuples( x, trans.columns ) )
    >  = List( tblsG2, x -> GetFusionMap( x, lib ) );
    true
    gap> ConstructModularGV4Tables( tblG, tblsG2, poss, lib );;
    #I  not all input tables for L2(81).(2x4) mod 3 available
    #I  not all input tables for L2(81).(2x4) mod 5 available
    #I  not all input tables for L2(81).(2x4) mod 41 available

6.10  The Character Table of O8+(3).22111

The construction of the character table of the group O8+(3).22111 is not as straightforward as the constructions shown in Section 6.2. Here we get 26 triples of actions on the tables of the three subgroups G.2i of index two, but only one of them leads to candidates for the desired character table. Specifically, we get 64 such candidates, in two equivalence classes w.r.t. permutation equivalence.
    gap> input:= [ "O8+(3)", "O8+(3).2_1",  "O8+(3).2_1'", "O8+(3).2_1''",
    >                                                  "O8+(3).(2^2)_{111}" ];;
    gap> tblG   := CharacterTable( input[1] );;
    gap> tblsG2 := List( input{ [ 2 .. 4 ] }, CharacterTable );;
    gap> name   := Concatenation( "new", input[5] );;
    gap> lib    := CharacterTable( input[5] );;
    gap> poss   := ConstructOrdinaryGV4Table( tblG, tblsG2, name, lib );;
    #I  newO8+(3).(2^2)_{111}: 2 equivalence classes
    gap> Length( poss );
    64
    gap> reps:= RepresentativesCharacterTables( poss );;
    gap> Length( reps );
    2

The two candidate tables differ only in four irreducible characters involving irrationalities on the classes of element order 28. All three subgroups G.2i contain elements of order 28 that do not lie in the simple group G; these classes are roots of the same (unique) class of element order 7. The centralizer C of an order 7 element in G.22 has order 112 = 24 ·7, the intersection of C with G has the structure 22 ×7 since G contains three classes of cyclic subgroups of the order 14, and each of the intersections of C with one of the subgroups G.2i has the structure 2 ×4 ×7, so the structure of C is 42 ×7 ≅ 4 ×28.
    gap> t:= reps[1].table;;
    gap> ord7:= Filtered( [ 1 .. NrConjugacyClasses( t ) ],                        
    >               i -> OrdersClassRepresentatives( t )[i] = 7 );
    [ 37 ]
    gap> SizesCentralizers( t ){ ord7 };
    [ 112 ]
    gap> ord28:= Filtered( [ 1 .. NrConjugacyClasses( t ) ],
    >               i -> OrdersClassRepresentatives( t )[i] = 28 );
    [ 112, 113, 114, 115, 161, 162, 163, 164, 210, 211, 212, 213 ]
    gap> List( reps[1].G2fusGV4, x -> Intersection( ord28, x ) );
    [ [ 112, 113, 114, 115 ], [ 161, 162, 163, 164 ], [ 210, 211, 212, 213 ] ]
    gap> sub:= CharacterTable( "Cyclic", 28 ) * CharacterTable( "Cyclic", 4 );;
    gap> List( reps, x -> Length( PossibleClassFusions( sub, x.table ) ) );
    [ 0, 96 ]

It turns out that only one of the two candidate tables admits a class fusion from the character table of C, thus we have determined the ordinary character table of O8+(3).22111.
    gap> trans:= TransformingPermutationsCharacterTables( reps[2].table, lib );;
    gap> IsRecord( trans );
    true
    gap> List( reps[2].G2fusGV4, x -> OnTuples( x, trans.columns ) )
    >  = List( tblsG2, x -> GetFusionMap( x, lib ) );
    true
    gap> ConstructModularGV4Tables( tblG, tblsG2, poss, lib );;
    #I  not all input tables for O8+(3).(2^2)_{111} mod 3 available

So also the p-modular tables of O8+(3).22111 can be computed this way, provided that the p-modular tables of the index 2 subgroups are available.

7  Examples for the Type 22.G

We compute the character table of a group of the type 22.G from the character tables of the three factor groups of the type 2.G, using the function PossibleCharacterTablesOfTypeV4G, see Section 3.4.

7.1  The Character Table of 22.Sz(8)

The three central involutions in 22.Sz(8) are permuted cyclicly by an outer automorphism α of order three. In order to apply PossibleCharacterTablesOfTypeV4G, we need the character table of the group 2.Sz(8) and the action on the classes of Sz(8) that is induced by α.
The ordinary character table of G = Sz(8) admits exactly five table automorphisms of order dividing 3. Each of these possibilities leads to exactly one possible character table of 22.G, and the five tables are permutation equivalent. From this point of view, we need not know which of the table automorphisms are induced by outer group automorphisms of G.
    gap> t:= CharacterTable( "Sz(8)" );;
    gap> 2t:= CharacterTable( "2.Sz(8)" );;
    gap> aut:= AutomorphismsOfTable( t );;
    gap> elms:= Set( List( Filtered( aut, x -> Order( x ) in [ 1, 3 ] ),           
    >                      SmallestGeneratorPerm ) );
    [ (), (9,10,11), (6,7,8), (6,7,8)(9,10,11), (6,7,8)(9,11,10) ]
    gap> poss:= List( elms,                                         
    >       pi -> PossibleCharacterTablesOfTypeV4G( t, 2t, pi, "2^2.Sz(8)" ) );
    [ [ CharacterTable( "2^2.Sz(8)" ) ], [ CharacterTable( "2^2.Sz(8)" ) ], 
      [ CharacterTable( "2^2.Sz(8)" ) ], [ CharacterTable( "2^2.Sz(8)" ) ], 
      [ CharacterTable( "2^2.Sz(8)" ) ] ]
    gap> reps:= RepresentativesCharacterTables( Concatenation( poss ) );
    [ CharacterTable( "2^2.Sz(8)" ) ]

The tables coincide with the one that is stored in the GAP library.
    gap> IsRecord( TransformingPermutationsCharacterTables( reps[1],
    >        CharacterTable( "2^2.Sz(8)" ) ) );
    true

The computation of the p-modular character table of 22.G from the p-modular character table of 2.G and the three factor fusions from 22.G to 2.G is straightforward, as is stated in Section 3.4. The three fusions are stored on the tables returned by PossibleCharacterTablesOfTypeV4G.
    gap> GetFusionMap( poss[1][1], 2t, "1" );
    [ 1, 1, 2, 2, 3, 4, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 
      13, 14, 14, 15, 15, 16, 16, 17, 17, 18, 18, 19, 19 ]
    gap> GetFusionMap( poss[1][1], 2t, "2" );
    [ 1, 2, 1, 2, 3, 4, 5, 6, 7, 6, 7, 8, 9, 8, 9, 10, 11, 10, 11, 12, 13, 12, 
      13, 14, 15, 14, 15, 16, 17, 16, 17, 18, 19, 18, 19 ]
    gap> GetFusionMap( poss[1][1], 2t, "3" );
    [ 1, 2, 2, 1, 3, 4, 5, 6, 7, 7, 6, 8, 9, 9, 8, 10, 11, 11, 10, 12, 13, 13, 
      12, 14, 15, 15, 14, 16, 17, 17, 16, 18, 19, 19, 18 ]

The GAP library function BrauerTableOfTypeV4G can be used to derive Brauer tables of 22.G. We have to compute the p-modular tables for prime divisors p of |G|, that is, for p ∈ { 2, 5, 7, 13 }.
    gap> Set( Factors( Size( t ) ) );
    [ 2, 5, 7, 13 ]

Clearly p = 2 is uninteresting from this point of view because the 2-modular table of 22.G can be identified with the 2-modular table of G.
For each of the five ordinary tables (corresponding to the five possible table automorphisms of G) constructed above, we get one candidate of a 5-modular table. However, these tables are not all equivalent. There are two equivalence classes, and one of the two possibilities is inconsistent in the sense that not all tensor products of irreducibles decompose into irreducibles.
    gap> cand:= List( poss, l -> BrauerTableOfTypeV4G( l[1], 2t mod 5,
    >      ConstructionInfoCharacterTable( l[1] )[3] ) );
    [ BrauerTable( "2^2.Sz(8)", 5 ), BrauerTable( "2^2.Sz(8)", 5 ), 
      BrauerTable( "2^2.Sz(8)", 5 ), BrauerTable( "2^2.Sz(8)", 5 ), 
      BrauerTable( "2^2.Sz(8)", 5 ) ]
    gap> Length( RepresentativesCharacterTables( cand ) );
    2
    gap> List( cand, CTblLib.Test.TensorDecomposition );
    [ false, true, false, true, true ]
    gap> Length( RepresentativesCharacterTables( cand{ [ 2, 4, 5 ] } ) );
    1
    gap> IsRecord( TransformingPermutationsCharacterTables( cand[2],
    >        CharacterTable( "2^2.Sz(8)" ) mod 5 ) );
    true

This implies that only those table automorphisms of G can be induced by an outer group automorphism that move the classes of element order 13.
The 7-modular table of 22.G is uniquely determined, independent of the choice of the table automorphism of G.
    gap> cand:= List( poss, l -> BrauerTableOfTypeV4G( l[1], 2t mod 7,
    >      ConstructionInfoCharacterTable( l[1] )[3] ) );
    [ BrauerTable( "2^2.Sz(8)", 7 ), BrauerTable( "2^2.Sz(8)", 7 ), 
      BrauerTable( "2^2.Sz(8)", 7 ), BrauerTable( "2^2.Sz(8)", 7 ), 
      BrauerTable( "2^2.Sz(8)", 7 ) ]
    gap> Length( RepresentativesCharacterTables( cand ) );
    1
    gap> IsRecord( TransformingPermutationsCharacterTables( cand[1],      
    >        CharacterTable( "2^2.Sz(8)" ) mod 7 ) );
    true

We get two candidates for the 13-modular table of 22.G, also if we consider only the three admissible table automorphisms.
    gap> elms:= elms{ [ 2, 4, 5 ] };
    [ (9,10,11), (6,7,8)(9,10,11), (6,7,8)(9,11,10) ]
    gap> poss:= poss{ [ 2, 4, 5 ] };;                                     
    gap> cand:= List( poss, l -> BrauerTableOfTypeV4G( l[1], 2t mod 13,
    >      ConstructionInfoCharacterTable( l[1] )[3] ) );
    [ BrauerTable( "2^2.Sz(8)", 13 ), BrauerTable( "2^2.Sz(8)", 13 ), 
      BrauerTable( "2^2.Sz(8)", 13 ) ]
    gap> Length( RepresentativesCharacterTables( cand ) );
    2
    gap> List( cand, CTblLib.Test.TensorDecomposition );                      
    [ true, true, true ]

The action of the outer automorphism of order three of G can be read off from the 2-modular table of G. Note that the ordinary and the 5-modular character table of G possess two independent table automorphisms of order three, whereas the group of table automorphisms of the 2-modular table has order three. (The reason is that the irrational values on the classes of the element orders 7 and 13 appear in the same irreducible 2-modular Brauer characters.)
    gap> mod2:= CharacterTable( "Sz(8)" ) mod 2;
    BrauerTable( "Sz(8)", 2 )
    gap> AutomorphismsOfTable( mod2 );
    Group([ (3,4,5)(6,7,8) ])
    gap> OrdersClassRepresentatives( mod2 );
    [ 1, 5, 7, 7, 7, 13, 13, 13 ]

This means that the first candidate is ruled out; this determines the 13-modular character table of 22.G.
    gap> Length( RepresentativesCharacterTables( cand{ [ 2, 3 ] } ) );
    1
    gap> IsRecord( TransformingPermutationsCharacterTables( cand[2],
    >        CharacterTable( "2^2.Sz(8)" ) mod 13 ) );
    true

7.2   ATLAS Tables of the Type 22.G (September 2005)

Besides 22.Sz(8) (cf. Section 7.1), 22.O8+(3) (cf. Section 7.3), and certain central extensions of L3(4) (cf. Section 7.4), the following examples of central extensions of nearly simple ATLAS groups G by a Klein four group occur.
    gap> listV4G:= [
    >      [ "2^2.L3(4)",         "2.L3(4)",     "L3(4)"       ],
    >      [ "2^2.L3(4).2_1",     "2.L3(4).2_1", "L3(4).2_1"   ],
    >      [ "(2^2x3).L3(4)",     "6.L3(4)",     "3.L3(4)"     ],
    >      [ "(2^2x3).L3(4).2_1", "6.L3(4).2_1", "3.L3(4).2_1" ],
    >      [ "2^2.O8+(2)",        "2.O8+(2)",    "O8+(2)"      ],
    >      [ "2^2.U6(2)",         "2.U6(2)",     "U6(2)"       ],
    >      [ "(2^2x3).U6(2)",     "6.U6(2)",     "3.U6(2)"     ],
    >      [ "2^2.2E6(2)",        "2.2E6(2)",    "2E6(2)"      ],
    >      [ "(2^2x3).2E6(2)",    "6.2E6(2)",    "3.2E6(2)"    ],
    > ];;

(For the tables of (22 ×3).G, with G one of L3(4), U6(2), or 2E6(2), we could alternatively use the tables of 22.G and 3.G, and the construction described in [Bre].)
The function for computing the candidates for the ordinary character tables is similar to the one from Section 6.2.
    gap> ConstructOrdinaryV4GTable:= function( tblG, tbl2G, name, lib )
    >      local ord3, nam, poss, reps, trans;
    > 
    >      # Compute the possible actions for the ordinary tables.
    >      ord3:= Set( List( Filtered( AutomorphismsOfTable( tblG ),
    >                                  x -> Order( x ) = 3 ),
    >                        SmallestGeneratorPerm ) );
    >      if 1 < Length( ord3 ) then
    >        Print( "#I  ", name,
    >               ": the action of the automorphism is not unique" );
    >      fi;
    >      # Compute the possible ordinary tables for the given actions.
    >      nam:= Concatenation( "new", name );
    >      poss:= Concatenation( List( ord3, pi ->
    >             PossibleCharacterTablesOfTypeV4G( tblG, tbl2G, pi, nam ) ) );
    >      # Test the possibilities for permutation equivalence.
    >      reps:= RepresentativesCharacterTables( poss );
    >      if 1 < Length( reps ) then
    >        Print( "#I  ", name, ": ", Length( reps ),
    >               " equivalence classes\n" );
    >      elif Length( reps ) = 0 then
    >        Print( "#E  ", name, ": no solution\n" );
    >      else
    >        # Compare the computed table with the library table.
    >        if not IsCharacterTable( lib ) then
    >          Print( "#I  no library table for ", name, "\n" );
    >          PrintToLib( name, poss[1].table );
    >        else
    >          trans:= TransformingPermutationsCharacterTables( reps[1], lib );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " differ\n" );
    >          fi;
    >        fi;
    >      fi;
    >      return poss;
    >    end;;

Concerning the Brauer tables, the same ambiguity problem may occur as in Section 6.2: Some candidates for the ordinary table may be excluded due to information provided by some p-modular table, see Section 7.1 for an easy example. Our strategy is analogous to the one used in Section 6.2.
    gap> ConstructModularV4GTables:= function( tblG, tbl2G, ordposs,
    >                                          ordlibtblV4G )
    >      local name, modposs, primes, checkordinary, i, p, tmodp, 2tmodp, aut,
    >            poss, modlib, trans, reps;
    > 
    >      if not IsCharacterTable( ordlibtblV4G ) then
    >        Print( "#I  no ordinary library table ...\n" );
    >        return [];
    >      fi;
    >      name:= Identifier( ordlibtblV4G );
    >      modposs:= [];
    >      primes:= Set( Factors( Size( tblG ) ) );
    >      ordposs:= ShallowCopy( ordposs );
    >      checkordinary:= false;
    >      for i in [ 1 .. Length( ordposs ) ] do
    >        modposs[i]:= [];
    >        for p in primes do
    >          tmodp := tblG  mod p;
    >          2tmodp:= tbl2G mod p;
    >          if IsCharacterTable( tmodp ) and IsCharacterTable( 2tmodp ) then
    >            aut:= ConstructionInfoCharacterTable( ordposs[i] )[3];
    >            poss:= BrauerTableOfTypeV4G( ordposs[i], 2tmodp, aut );
    >            if CTblLib.Test.TensorDecomposition( poss, false ) = false then
    >              Print( "#I  excluded cand. ", i, " (out of ",
    >                     Length( ordposs ), ") for ", name, " by ", p,
    >                     "-mod. table\n" );
    >              Unbind( ordposs[i] );
    >              Unbind( modposs[i] );
    >              checkordinary:= true;
    >              break;
    >            fi;
    >            Add( modposs[i], poss );
    >          else
    >            Print( "#I  not all input tables for ", name, " mod ", p,
    >                   " available\n" );
    >            primes:= Difference( primes, [ p ] );
    >          fi;
    >        od;
    >        if IsBound( modposs[i] ) then
    >          # Compare the computed Brauer tables with the library tables.
    >          for poss in modposs[i] do
    >            p:= UnderlyingCharacteristic( poss );
    >            modlib:= ordlibtblV4G mod p;
    >            if IsCharacterTable( modlib ) then
    >              trans:= TransformingPermutationsCharacterTables(
    >                          poss, modlib );
    >              if not IsRecord( trans ) then
    >                Print( "#E  computed table and library table for ",
    >                       name, " mod ", p, " differ\n" );
    >              fi;
    >            else
    >              Print( "#I  no library table for ",
    >                     name, " mod ", p, "\n" );
    >              PrintToLib( name, poss );
    >            fi;
    >          od;
    >        fi;
    >      od;
    >      if checkordinary then
    >        # Test whether the ordinary table is admissible.
    >        ordposs:= Compacted( ordposs );
    >        modposs:= Compacted( modposs );
    >        reps:= RepresentativesCharacterTables( ordposs );
    >        if 1 < Length( reps ) then
    >          Print( "#I  ", name, ": ", Length( reps ),
    >                 " equivalence classes (ord. table)\n" );
    >        elif Length( reps ) = 0 then
    >          Print( "#E  ", name, ": no solution (ord. table)\n" );
    >        else
    >          # Compare the computed table with the library table.
    >          trans:= TransformingPermutationsCharacterTables( reps[1],
    >                      ordlibtblV4G );
    >          if not IsRecord( trans ) then
    >            Print( "#E  computed table and library table for ", name,
    >                   " differ\n" );
    >          fi;
    >        fi;
    >      fi;
    >      # Test the uniqueness of the Brauer tables.
    >      for poss in TransposedMat( modposs ) do
    >        reps:= RepresentativesCharacterTables( poss );
    >        if Length( reps ) <> 1 then
    >          Print( "#I  ", name, ": ", Length( reps ), " candidates for the ",
    >                 UnderlyingCharacteristic( reps[1] ), "-modular table\n" );
    >        fi;
    >      od;
    >      return rec( ordinary:= ordposs, modular:= modposs );
    >    end;;

In our examples, the action of the outer automorphism of order three on the classes of G turns out to be uniquely determined by the table automorphisms of the character table of G.
    gap> for input in listV4G do
    >      tblG  := CharacterTable( input[3] );
    >      tbl2G := CharacterTable( input[2] );
    >      lib   := CharacterTable( input[1] );
    >      poss  := ConstructOrdinaryV4GTable( tblG, tbl2G, input[1], lib );
    >      ConstructModularV4GTables( tblG, tbl2G, poss, lib );
    >    od;
    #I  excluded cand. 1 (out of 16) for 2^2.L3(4).2_1 by 7-mod. table
    #I  excluded cand. 2 (out of 16) for 2^2.L3(4).2_1 by 7-mod. table
    #I  excluded cand. 7 (out of 16) for 2^2.L3(4).2_1 by 7-mod. table
    #I  excluded cand. 10 (out of 16) for 2^2.L3(4).2_1 by 7-mod. table
    #I  excluded cand. 15 (out of 16) for 2^2.L3(4).2_1 by 7-mod. table
    #I  excluded cand. 16 (out of 16) for 2^2.L3(4).2_1 by 7-mod. table
    #I  excluded cand. 1 (out of 16) for (2^2x3).L3(4).2_1 by 7-mod. table
    #I  excluded cand. 2 (out of 16) for (2^2x3).L3(4).2_1 by 7-mod. table
    #I  excluded cand. 7 (out of 16) for (2^2x3).L3(4).2_1 by 7-mod. table
    #I  excluded cand. 10 (out of 16) for (2^2x3).L3(4).2_1 by 7-mod. table
    #I  excluded cand. 15 (out of 16) for (2^2x3).L3(4).2_1 by 7-mod. table
    #I  excluded cand. 16 (out of 16) for (2^2x3).L3(4).2_1 by 7-mod. table
    #I  not all input tables for 2^2.2E6(2) mod 2 available
    #I  not all input tables for 2^2.2E6(2) mod 3 available
    #I  not all input tables for 2^2.2E6(2) mod 5 available
    #I  not all input tables for 2^2.2E6(2) mod 7 available
    #I  not all input tables for (2^2x3).2E6(2) mod 2 available
    #I  not all input tables for (2^2x3).2E6(2) mod 3 available
    #I  not all input tables for (2^2x3).2E6(2) mod 5 available
    #I  not all input tables for (2^2x3).2E6(2) mod 7 available
    #I  not all input tables for (2^2x3).2E6(2) mod 11 available
    #I  not all input tables for (2^2x3).2E6(2) mod 13 available
    #I  not all input tables for (2^2x3).2E6(2) mod 17 available
    #I  not all input tables for (2^2x3).2E6(2) mod 19 available

7.3  The Character Table of 22.O8+(3) (March 2009)

When one tries to construct the character table of the central extensions of G = O8+(3) by a Klein four group, in the same way as in Section 7.2, one notices that the order three automorphism that relates the three central extensions of G by an involution is not uniquely determined.
    gap> entry:= [ "2^2.O8+(3)", "2.O8+(3)", "O8+(3)" ];;
    gap> tblG:= CharacterTable( entry[3] );;
    gap> aut:= AutomorphismsOfTable( tblG );;
    gap> ord3:= Set( List( Filtered( aut, x -> Order( x ) = 3 ),
    >                      SmallestGeneratorPerm ) );;
    gap> Length( ord3 );
    4

However, the table candidates one gets from the four possible automorphisms turn out to be all equivalent, hence the character table of 22.O8+(3) can be constructed as follows.
    gap> poss:= [];;
    gap> tbl2G:= CharacterTable( entry[2] );
    CharacterTable( "2.O8+(3)" )
    gap> for pi in ord3 do
    >   Append( poss,
    >           PossibleCharacterTablesOfTypeV4G( tblG, tbl2G, pi, entry[1] ) );
    > od;
    gap> Length( poss );
    32
    gap> poss:= RepresentativesCharacterTables( poss );;
    gap> Length( poss );
    1

The computed table coincides with the library table.
    gap> lib:= CharacterTable( entry[1] );;
    gap> if TransformingPermutationsCharacterTables( poss[1], lib ) = fail then
    >      Print( "#E  differences for ", entry[1], "\n" );
    >    fi;

7.4  The Character Table of the Schur Cover of L3(4) (September 2005)

The Schur cover of G = L3(4) has the structure (42 ×3).L3(4). Following [CCN+85,p. 23], we regard the multiplier of G as
M = 〈a, b, c, d | [a,b] = [a,c] = [a,d] = [b,c] = [b,d] = [c,d] = a4 = b4 = c4 = d3 = abc 〉,
and we will consider the automorphism α of M.G that acts as (a,b,c)(d) on M.
The subgroup lattice of the subgroup 〈a, b, c 〉 = 〈a, b 〉 ≅ 42 of M looks as follows. (The subgroup in the centre of the picture is the Klein four group 〈a2, b2, c2 〉 = 〈a2, b2 〉.)
ctblcons17.png
(The symmetry w.r.t. α would be reflected better in a three dimensional model, with 〈a, b 〉, 〈a2, b2 〉, and the trivial subgroup on a vertical symmetry axis, and with the remaining subgroups on three circles such that α induces a rotation.)
We have (M / 〈a 〉).G ≅ (M / 〈b 〉).G ≅ (M / 〈c 〉).G ≅ 122.G and (M / 〈a b2 〉).G ≅ (M / 〈b c2 〉).G ≅ (M / 〈c a2 〉).G ≅ 121.G. This is because the action of G.22 fixes a, and swaps b and c; so b is inverted modulo 〈a 〉 but fixed modulo 〈a b2 〉, and the normal subgroup of order four in 42.G.22 is central but that in 41.G.22 is not central.
The constructions of the character tables of 42.G and (42 ×3).G are essentially the same. We start with the table of 42.G. It can be regarded as a central extension H = V.22.G of 22.G by a Klein four group V. The three subgroups of order two in V are cyclicly permuted by the automorphism of M / 〈d 〉 induced by α, so the three factors by these subgroups are isomorphic groups F, say, with the structure (2 ×4).G.
The group F itself is a central extension of 2.G by a Klein four group, but in this case the three factor groups by the order two subgroups of the Klein four group are nonisomorphic groups, of the types 41.G, 42.G, and 22.G, respectively. The GAP function PossibleCharacterTablesOfTypeV4G can be used to construct the character table of F from the three factors. Note that in this case, no information about table automorphisms is required.
    gap> tblG:= CharacterTable( "2.L3(4)" );;
    gap> tbls2G:= List( [ "4_1.L3(4)", "4_2.L3(4)", "2^2.L3(4)"],
    >                   CharacterTable );;
    gap> poss:= PossibleCharacterTablesOfTypeV4G( tblG, tbls2G, "(2x4).L3(4)" );;
    gap> Length( poss );
    2
    gap> reps:= RepresentativesCharacterTables( poss );
    [ CharacterTable( "(2x4).L3(4)" ) ]
    gap> lib:= CharacterTable( "(2x4).L3(4)" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( reps[1], lib ) );
    true

In the second step, we construct the table of 42.G from that of (2 ×4).G and the table automorphism of 22.G that is induced by α; it turns out that the group of table automorphisms of 22.G contains a unique subgroup of order three.
    gap> tblG:= tbls2G[3];
    CharacterTable( "2^2.L3(4)" )
    gap> tbl2G:= lib;       
    CharacterTable( "(2x4).L3(4)" )
    gap> aut:= AutomorphismsOfTable( tblG );;
    gap> ord3:= Set( List( Filtered( aut, x -> Order( x ) = 3 ),
    >                  SmallestGeneratorPerm ) );
    [ (2,3,4)(6,7,8)(10,11,12)(13,15,17)(14,16,18)(20,21,22)(24,25,26)(28,29,
        30)(32,33,34) ]
    gap> pi:= ord3[1];;
    gap> poss:= PossibleCharacterTablesOfTypeV4G( tblG, tbl2G, pi, "4^2.L3(4)" );;
    gap> Length( poss );
    4
    gap> reps:= RepresentativesCharacterTables( poss );        
    [ CharacterTable( "4^2.L3(4)" ) ]
    gap> lib:= CharacterTable( "4^2.L3(4)" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( reps[1], lib ) );
    true

With the same approach, we compute the table of (2 ×12).G = 22.6.G from the tables of the three nonisomorphic factor groups 121.G, 122.G, and (22 ×3).G, and we compute the table of (42 ×3).G = 22.(22 ×3).G from the three tables of the factor groups (2 ×12).G and the action induced by α.
    gap> tblG:= CharacterTable( "6.L3(4)" );;
    gap> tbls2G:= List( [ "12_1.L3(4)", "12_2.L3(4)", "(2^2x3).L3(4)"],            
    >                   CharacterTable );;
    gap> poss:= PossibleCharacterTablesOfTypeV4G( tblG, tbls2G, "(2x12).L3(4)" );;
    gap> Length( poss );
    2
    gap> reps:= RepresentativesCharacterTables( poss );
    [ CharacterTable( "(2x12).L3(4)" ) ]
    gap> lib:= CharacterTable( "(2x12).L3(4)" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( reps[1], lib ) );
    true
    gap> tblG:= CharacterTable( "(2^2x3).L3(4)" ); 
    CharacterTable( "(2^2x3).L3(4)" )
    gap> tbl2G:= CharacterTable( "(2x12).L3(4)" );
    CharacterTable( "(2x12).L3(4)" )
    gap> aut:= AutomorphismsOfTable( tblG );;
    gap> ord3:= Set( List( Filtered( aut, x -> Order( x ) = 3 ),
    >                  SmallestGeneratorPerm ) );
    [ (2,7,8)(3,4,10)(6,11,12)(14,19,20)(15,16,22)(18,23,24)(26,27,28)(29,35,
        41)(30,37,43)(31,39,45)(32,36,42)(33,38,44)(34,40,46)(48,53,54)(49,50,
        56)(52,57,58)(60,65,66)(61,62,68)(64,69,70)(72,77,78)(73,74,80)(76,81,
        82)(84,89,90)(85,86,92)(88,93,94) ]
    gap> pi:= ord3[1];;
    gap> poss:= PossibleCharacterTablesOfTypeV4G( tblG, tbl2G, pi,
    >                                             "(4^2x3).L3(4)" );;
    gap> Length( poss );
    4
    gap> reps:= RepresentativesCharacterTables( poss );
    [ CharacterTable( "(4^2x3).L3(4)" ) ]
    gap> lib:= CharacterTable( "(4^2x3).L3(4)" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( reps[1], lib ) );
    true

8  Examples of Extensions by p-singular Automorphisms

8.1  Some p-Modular Tables of Groups of the Type M.G.A

We show an alternative construction of p-modular tables of certain groups that have been met in Section 4.3. Each entry in the GAP list listMGA contains the Identifier values of character tables of groups of the types M.G, G, G.A, and M.G.A. For each entry with |A| = p, a prime integer, we fetch the p-modular table of G and the ordinary table of G.A, compute the action of G.A on the p-regular classes of G, and then compute the p-modular table of G.A. Analogously, we compute the p-modular table of M.G.A from the p-modular table of M.G and the ordinary table of M.G.A.
    gap> for input in listMGA do
    >      ordtblMG  := CharacterTable( input[1] );
    >      ordtblG   := CharacterTable( input[2] );
    >      ordtblGA  := CharacterTable( input[3] );
    >      ordtblMGA := CharacterTable( input[4] );
    >      p:= Size( ordtblGA ) / Size( ordtblG );
    >      if IsPrimeInt( p ) then
    >        modtblG:= ordtblG mod p;
    >        if modtblG <> fail then
    >          modtblGA := CharacterTableRegular( ordtblGA, p );
    >          SetIrr( modtblGA, IBrOfExtensionBySingularAutomorphism( modtblG,
    >                                ordtblGA ) );
    >          if TransformingPermutationsCharacterTables( modtblGA,
    >                 ordtblGA mod p ) = fail then
    >            Print( "#E  computed table and library table for ", input[3],
    >                   " mod ", p, " differ\n" );
    >          fi;
    >        fi;
    >        modtblMG:= ordtblMG mod p;
    >        if modtblMG <> fail then
    >          modtblMGA := CharacterTableRegular( ordtblMGA, p );
    >          SetIrr( modtblMGA, IBrOfExtensionBySingularAutomorphism( modtblMG,
    >                                 ordtblMGA ) );
    >          if TransformingPermutationsCharacterTables( modtblMGA,
    >                 ordtblMGA mod p ) = fail then
    >            Print( "#E  computed table and library table for ", input[4],
    >                   " mod ", p, " differ\n" );
    >          fi;
    >        fi;
    >      fi;
    >    od;

8.2  Some p-Modular Tables of Groups of the Type G.S3

We show an alternative construction of 2- and 3-modular tables of certain groups that have been met in Section 5.2. Each entry in the GAP list listGS3 contains the Identifier values of character tables of groups of the types G, G.2, G.3, and G.S3. For each entry, we fetch the 2-modular table of G and the ordinary table of G.2, compute the action of G.2 on the 2-regular classes of G, and then compute the 2-modular table of G.2. Analogously, we compute the 3-modular table of G.3 from the 3-modular table of G and the ordinary table of G.3, and we compute the 2-modular table of G.S3 from the 2-modular table of G.3 and the ordinary table of G.S3.
    gap> for input in listGS3 do
    >      modtblG:= CharacterTable( input[1] ) mod 2;
    >      if modtblG <> fail then
    >        ordtblG2 := CharacterTable( input[2] );
    >        modtblG2 := CharacterTableRegular( ordtblG2, 2 );
    >        SetIrr( modtblG2, IBrOfExtensionBySingularAutomorphism( modtblG,
    >                              ordtblG2 ) );
    >        if TransformingPermutationsCharacterTables( modtblG2,
    >               ordtblG2 mod 2 ) = fail then
    >          Print( "#E  computed table and library table for ", input[2],
    >                 " mod 2 differ\n" );
    >        fi;
    >      fi;
    >      modtblG:= CharacterTable( input[1] ) mod 3;
    >      if modtblG <> fail then
    >      ordtblG3 := CharacterTable( input[3] );
    >        modtblG3 := CharacterTableRegular( ordtblG3, 3 );
    >        SetIrr( modtblG3, IBrOfExtensionBySingularAutomorphism( modtblG,
    >                              ordtblG3 ) );
    >        if TransformingPermutationsCharacterTables( modtblG3,
    >               ordtblG3 mod 3 ) = fail then
    >          Print( "#E  computed table and library table for ", input[3],
    >                 " mod 3 differ\n" );
    >        fi;
    >      fi;
    >      modtblG3:= CharacterTable( input[3] ) mod 2;
    >      if modtblG3 <> fail then
    >        ordtblGS3 := CharacterTable( input[4] );
    >        modtblGS3 := CharacterTableRegular( ordtblGS3, 2 );
    >        SetIrr( modtblGS3, IBrOfExtensionBySingularAutomorphism( modtblG3,
    >                               ordtblGS3 ) );
    >        if TransformingPermutationsCharacterTables( modtblGS3,
    >               ordtblGS3 mod 2 ) = fail then
    >          Print( "#E  computed table and library table for ", input[4],
    >                 " mod 2 differ\n" );
    >        fi;
    >      fi;
    >    od;

8.3  2-Modular Tables of Groups of the Type G.22

We show an alternative construction of 2-modular tables of certain groups that have been met in Section 6.2. Each entry in the GAP list listGV4 contains the Identifier values of character tables of groups of the types G, G.21, G.22, G.23, and G.22. For each entry, we fetch the 2-modular table of G and the ordinary tables of the groups G.2i, and compute the 2-modular tables of G.2i; Then we compute from this modular table and the ordinary table of G.22 the 2-modular table of G.22.
    gap> for input in listGV4 do
    >      modtblG   := CharacterTable( input[1] ) mod 2;
    >      if modtblG <> fail then
    >        ordtblsG2 := List( input{ [ 2 .. 4 ] }, CharacterTable );
    >        ordtblGV4 := CharacterTable( input[5] );
    >        for tblG2 in ordtblsG2 do
    >          modtblG2:= CharacterTableRegular( tblG2, 2 );
    >          SetIrr( modtblG2, IBrOfExtensionBySingularAutomorphism( modtblG,
    >                                tblG2 ) );
    >          if TransformingPermutationsCharacterTables( modtblG2,
    >                 tblG2 mod 2 ) = fail then
    >            Print( "#E  computed table and library table for ",
    >                   Identifier( tblG2 ), " mod 2 differ\n" );
    >          fi;
    >          modtblGV4:= CharacterTableRegular( ordtblGV4, 2 );
    >          SetIrr( modtblGV4, IBrOfExtensionBySingularAutomorphism( modtblG2,
    >                                ordtblGV4 ) );
    >          if TransformingPermutationsCharacterTables( modtblGV4,
    >                 ordtblGV4 mod 2 ) = fail then
    >            Print( "#E  computed table and library table for ", input[5],
    >                   " mod 2 differ\n" );
    >          fi;
    >        od;
    >      fi;
    >    od;

9  Examples of Subdirect Products of Index Two

Typical examples of this construction are those maximal subgroups of alternating groups An that extend in the corresponding symmetric groups Sn to direct products of the structures Sm ×Sn−m, for 2 < m < n/2. Also certain subgroups of these maximal subgroups that have this structure can be interesting, see Section 4.2.

9.1  Certain Dihedral Groups as Subdirect Products of Index Two

Also dihedral groups of order 2 n with n divisible by at least two different primes have the required structure: Let n = n1 n2 with coprime n1, n2, and let the normal subgroups H1, H2 be cyclic subgroups of order n1 and n2, respectively, inside the cyclic subgroup of index two. Then the factors G/N1, G/N2 are themselves dihedral groups.
So an example (with n1 = 3 and n2 = 5) is the construction of the dihedral group D30 as a subdirect product of index two in the direct product D6 ×D10.
    gap> tblh1:= CharacterTable( "C3" );;
    gap> tblg1:= CharacterTable( "S3" );;
    gap> StoreFusion( tblh1, PossibleClassFusions( tblh1, tblg1 )[1], tblg1 );
    gap> tblh2:= CharacterTable( "C5" );;
    gap> tblg2:= CharacterTable( "D10" );;
    gap> StoreFusion( tblh2, PossibleClassFusions( tblh2, tblg2 )[1], tblg2 );
    gap> subdir:= CharacterTableOfIndexTwoSubdirectProduct( tblh1, tblg1,
    >                 tblh2, tblg2, "D30" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( subdir.table,
    >                  CharacterTable( "Dihedral", 30 ) ) );
    true

9.2  The Character Table of (D10 ×HN).2 < M (June 2008)

The sporadic simple Monster group contains maximal subgroups with the structure (D10 ×HN).2 (see [CCN+85,p. 234]), the factor group modulo D10 is the automorphism group HN.2 of HN, and the factor group modulo HN is the Frobenius group 5:4 of order 20.
    gap> tblh1:= CharacterTable( "D10" );;
    gap> tblg1:= CharacterTable( "5:4" );;
    gap> tblh2:= CharacterTable( "HN" );;
    gap> tblg2:= CharacterTable( "HN.2" );;
    gap> subdir:= CharacterTableOfIndexTwoSubdirectProduct( tblh1, tblg1,
    >                 tblh2, tblg2, "(D10xHN).2" );;
    gap> IsRecord( TransformingPermutationsCharacterTables( subdir.table,
    >                  CharacterTable( "(D10xHN).2" ) ) );
    true
    gap> m:= CharacterTable( "M" );;
    gap> fus:= PossibleClassFusions( subdir.table, m );;
    gap> Length( fus );
    16
    gap> Length( RepresentativesFusions( subdir.table, fus, m ) );
    1

An alternative construction is the one described in Section 3.1, as (D10 ×HN).2 = M.G.A with G = 2 ×HN, M.G = D10 ×HN, and G.A the subdirect product of HN.2 and a cyclic group of order four (which can be constructed as the isoclinic variant of 2 ×HN.2, see Section 2.4).
Here is this construction:
    gap> c2:= CharacterTable( "C2" );;
    gap> hn:= CharacterTable( "HN" );;
    gap> g:= c2 * hn;;
    gap> d10:= CharacterTable( "D10" );;
    gap> mg:= d10 * hn;;
    gap> nsg:= ClassPositionsOfNormalSubgroups( mg );
    [ [ 1 ], [ 1, 55 .. 109 ], [ 1, 55 .. 163 ], [ 1 .. 54 ], [ 1 .. 162 ], 
      [ 1 .. 216 ] ]
    gap> SizesConjugacyClasses( mg ){ nsg[2] };
    [ 1, 2, 2 ]
    gap> g:= mg / nsg[2];
    CharacterTable( "D10xHN/[ 1, 55, 109 ]" )
    gap> help:= c2 * CharacterTable( "HN.2" );
    CharacterTable( "C2xHN.2" )
    gap> ga:= CharacterTableIsoclinic( help ); 
    CharacterTable( "Isoclinic(C2xHN.2)" )
    gap> gfusga:= PossibleClassFusions( g, ga ); 
    [ [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 11, 12, 13, 14, 15, 16, 17, 18, 19, 
          20, 21, 22, 23, 23, 24, 25, 25, 26, 27, 28, 29, 30, 31, 32, 32, 33, 33, 
          34, 35, 36, 37, 37, 38, 39, 40, 40, 41, 42, 42, 43, 43, 44, 44, 79, 80, 
          81, 82, 83, 84, 85, 86, 87, 88, 89, 89, 90, 91, 92, 93, 94, 95, 96, 97, 
          98, 99, 100, 101, 101, 102, 103, 103, 104, 105, 106, 107, 108, 109, 
          110, 110, 111, 111, 112, 113, 114, 115, 115, 116, 117, 118, 118, 119, 
          120, 120, 121, 121, 122, 122 ], 
      [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 11, 12, 13, 14, 15, 16, 17, 18, 19, 
          20, 21, 22, 23, 23, 24, 25, 25, 26, 27, 28, 29, 30, 31, 32, 32, 33, 33, 
          35, 34, 36, 37, 37, 38, 39, 40, 40, 41, 42, 42, 43, 43, 44, 44, 79, 80, 
          81, 82, 83, 84, 85, 86, 87, 88, 89, 89, 90, 91, 92, 93, 94, 95, 96, 97, 
          98, 99, 100, 101, 101, 102, 103, 103, 104, 105, 106, 107, 108, 109, 
          110, 110, 111, 111, 113, 112, 114, 115, 115, 116, 117, 118, 118, 119, 
          120, 120, 121, 121, 122, 122 ] ]
    gap> StoreFusion( g, gfusga[1], ga );
    gap> acts:= PossibleActionsForTypeMGA( mg, g, ga );;
    gap> Length( acts );
    1
    gap> poss:= PossibleCharacterTablesOfTypeMGA( mg, g, ga, acts[1],       
    >               "(D10xHN).2" );;
    gap> Length( poss );
    1
    gap> IsRecord( TransformingPermutationsCharacterTables( poss[1].table,
    >                  CharacterTable( "(D10xHN).2" ) ) );
    true

References

[Bre]
T. Breuer, Constructing character tables of central extensions in GAP, http://www.math.rwth-aachen.de/ ~Thomas.Breuer/ ctbllib/doc/ ctocenex.pdf.
[Bre12]


, The GAP Character Table Library, Version 1.2, http://www.math.rwth-aachen.de/ ~Thomas.Breuer/ ctbllib, Mar 2012, GAP package.
[CCN+85]
J. H. Conway, R. T. Curtis, S. P. Norton, R. A. Parker, and R. A. Wilson, Atlas of finite groups, Oxford University Press, Eynsham, 1985, Maximal subgroups and ordinary characters for simple groups, With computational assistance from J. G. Thackray. MR 827219 (88g:20025)
[Fei82]
W. Feit, The representation theory of finite groups, North-Holland Mathematical Library, vol. 25, North-Holland Publishing Co., 1982, xiv+502 pp., ISBN 0-444-86155-6. MR 661045 (83g:20001)
[Gag86]
S. M. Gagola, Jr., Formal character tables, Michigan Math. J. 33 (1986), no. 1, 3-10. MR 817904 (86k:20010)
[GAP04]
The GAP Group, GAP-Groups, Algorithms, and Programming, Version 4.4, 2004, http://www.gap-system.org.
[HL94]
G. Hiss and K. Lux, The 5-modular characters of the sporadic simple Fischer groups Fi22 and Fi23, Comm. Algebra 22 (1994), no. 9, 3563-3590, With an appendix by Thomas Breuer. MR 1278806 (95e:20020)
[Isa76]
I. M. Isaacs, Character theory of finite groups, Academic Press [Harcourt Brace Jovanovich Publishers], New York, 1976, Pure and Applied Mathematics, No. 69. MR 0460423 (57 #417)



File translated from TEX by TTH, version 3.59.
On 5 May 2012, 09:04.