  
  [1X11 [33X[0;0YExamples[133X[101X
  
  [33X[0;0YTo  actually  run an orbit enumeration by suborbits, we have to collect some
  insight  into  the  structure  of the group under consideration and into its
  representation  theory.  In  general, preparing the input data is more of an
  art than a science. The mathematical details are described in [MNW07].[133X
  
  [33X[0;0YIn   Section   [14X11.1[114X  we  present  a  small  example  of  the  usage  of  the
  orbit-by-suborbit machinery. We use the sporadic simple Mathieu group [22XM_{11}[122X
  acting projectively on its irreducible module of dimension 24 over the field
  with 3 elements.[133X
  
  [33X[0;0YIn   Section   [14X11.2[114X   we  present  another  example  of  the  usage  of  the
  orbit-by-suborbit programs. In this example we determine 35 of the 36 double
  coset  representatives  of  the  sporadic  simple Fischer group [22XFi_{23}[122X with
  respect to its seventh maximal subgroup.[133X
  
  [33X[0;0YIn   Section  [14X11.3[114X  we  present  a  bigger  example  of  the  usage  of  the
  orbit-by-suborbit  machinery.  In  this  example  the  orbit  lengths of the
  sporadic  simple Conway group [22XCo_{1}[122X acting in in its irreducible projective
  representation   over  the  field  with  [22X5[122X  elements  in  dimension  [22X24[122X  are
  determined,  which  were previously unknown. These orbit lengths were needed
  to rule out a case in [Mal06].[133X
  
  [33X[0;0YIn  Section  [14X11.4[114X  we present as an extended worked example how to enumerate
  the  smallest non-trivial orbit of the sporadic simple Baby Monster group [22XB[122X.
  We  give a log of a [5XGAP[105X session with explanations in between, being intended
  to  illustrate  a few of the tools which are available in the [5Xorb[105X package as
  well as in related packages. Actually, the [5Xorb[105X package has also been applied
  to  two  much  larger  permutation actions of [22XB[122X, namely its action on its 2B
  involutions,  having  degree [22X≈ 1.2⋅ 10^13[122X, and its action on the cosets of a
  maximal  subgroup  isomorphic  to  [22XFi_23[122X,  having  degree  [22X≈ 1.0⋅ 10^15[122X; for
  details see [M\t08] and [MNW07], respectively.[133X
  
  [33X[0;0YNote  that for all this to work you have to acquire and install the packages
  [5XIO[105X,  [5Xcvec[105X,  and  [5Xatlasrep[105X,  and  for  Section [14X11.4[114X you additionally need the
  packages [5Xchop[105X and [5Xgenss[105X.[133X
  
  
  [1X11.1 [33X[0;0YThe Mathieu group [22XM_{11}[122X[101X[1X acting in dimension [22X24[122X[101X[1X[133X[101X
  
  [33X[0;0YThe  example  in  this  section  is  very  small  but  our intention is that
  everything can still be analysed and looked at more or less by hand. We want
  to  enumerate  orbits of the Mathieu group [22XM_{11}[122X acting projectively on its
  irreducible  module  of dimension 24 over the field with 3 elements. All the
  files for this example are located in the [11Xexamples/m11PF3d24[111X subdirectory of
  the [5Xorb[105X package. Then you simply run the example in the following way:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/m11PF3d24/M11OrbitOnPF3d24.g");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27Xo := OrbitBySuborbit(setup,v,3,3,2,100);[127X[104X
    [4X[28X...[128X[104X
    [4X[28X#I  OrbitBySuborbit found 100% of a U3-orbit of size 7 920[128X[104X
    [4X[28X...[128X[104X
  [4X[32X[104X
  
  [33X[0;0YEverything  works  instantly as it would have without the orbit-by-suborbits
  method.  (Depending  on  whether  the  matrix and permutation generators for
  [22XM_{11}[122X are already stored locally, some time might be needed to fetch them.)
  The  details  of  this computation can be directly read off from the code in
  the file [11XM11OrbitOnPF3d24.g[111X:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[28XLoadPackage("orb");[128X[104X
    [4X[28XLoadPackage("io");[128X[104X
    [4X[28XLoadPackage("cvec");[128X[104X
    [4X[28XLoadPackage("atlasrep");[128X[104X
    [4X[28X[128X[104X
    [4X[28XSetInfoLevel(InfoOrb,2);[128X[104X
    [4X[28Xpgens := AtlasGenerators([ "M11", [ "M11G1-p11B0.m1", "M11G1-p11B0.m2" ], 1, 11 ]).generators;[128X[104X
    [4X[28X[128X[104X
    [4X[28Xgens := AtlasGenerators([ "M11", [ "M11G1-f3r24B0.m1", "M11G1-f3r24B0.m2" ], 1, 3 ]).generators;[128X[104X
    [4X[28Xcgens := List(gens,CMat);[128X[104X
    [4X[28Xbasech := CVEC_ReadMatFromFile(Filename(DirectoriesPackageLibrary("orb",""),[128X[104X
    [4X[28X          "examples/m11PF3d24/m11basech.cmat"));[128X[104X
    [4X[28Xbasechi := basech^-1;[128X[104X
    [4X[28Xcgens := List(cgens,x->basech*x*basechi);[128X[104X
    [4X[28X[128X[104X
    [4X[28XReadPackage("orb","examples/m11PF3d24/m11slps.g");[128X[104X
    [4X[28Xpgu2 := ResultOfStraightLineProgram(s2,pgens);[128X[104X
    [4X[28Xpgu1 := ResultOfStraightLineProgram(s1,pgu2);[128X[104X
    [4X[28Xcu2 := ResultOfStraightLineProgram(s2,cgens);[128X[104X
    [4X[28Xcu1 := ResultOfStraightLineProgram(s1,cu2);[128X[104X
    [4X[28X[128X[104X
    [4X[28Xsetup := OrbitBySuborbitBootstrapForLines([128X[104X
    [4X[28X         [cu1,cu2,cgens],[pgu1,pgu2,pgens],[20,720,7920],[5,11],rec());[128X[104X
    [4X[28Xsetup!.stabchainrandom := 900;[128X[104X
    [4X[28X[128X[104X
    [4X[28Xv := ZeroMutable(cgens[1][1]);[128X[104X
    [4X[28XRandomize(v);[128X[104X
    [4X[28XORB_NormalizeVector(v); [128X[104X
    [4X[28X[128X[104X
    [4X[28XPrint("Now do\n  o := OrbitBySuborbit(setup,v,3,3,2,100);\n");[128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  are using two helper subgroups [22XU_1 < U_2 < M_11[122X, where [22XU_2≅ A_6.2[122X is the
  largest  maximal  subgroup  of  [22XM_11[122X,  having  order  [22X720[122X, and [22XU_2≅ 5:4[122X is a
  maximal  subgroup  of  [22XU_2[122X of order [22X20[122X, see [CCN+85] or the [5XCTblLib[105X package.
  The quotient spaces we use for the helper subgroups have dimensions [22X5[122X and [22X11[122X
  respectively.  Straight  line  programs  to compute generators of the helper
  subgroups in terms of the given generators of [22XM_11[122X, and an appropriate basis
  exhibiting  the quotients, have already been computed, and are stored in the
  files  [11Xm11slps.g[111X  and [11Xm11basech.cmat[111X, respectively. (In Section [14X11.4[114X we show
  in  detail  how  such straight line programs and suitable bases can be found
  using   the   tools   available   in   in  the  [5Xorb[105X  package.)  The  command
  [11XOrbitBySuborbitBootstrapForLines[111X   invokes   the   precomputation,   and  in
  particular says that we want to use projective action.[133X
  
  
  [1X11.2 [33X[0;0YThe Fischer group [22XFi_{23}[122X[101X[1X acting in dimension [22X1494[122X[101X[1X[133X[101X
  
  [33X[0;0YThe  example  in this section shows how to compute 35 of the 36 double coset
  representatives  of  the  Fischer  group [22XFi_{23}[122X with respect to its seventh
  maximal  subgroup  [22XH≅  3_+^{1+8}.2_-^{1+6}.3_+^{1+2}.2S_4[122X,  which  has order
  [22X3265173504≈  3.2⋅  10^9[122X  and index [22X[Fi_{23}: H]=1252451200 ≈ 1.3 ⋅ 10^9[122X, see
  [CCN+85]  or the [5XCTblLib[105X package. All the files for this example are located
  in  the  [11Xexamples/fi23m7[111X subdirectory of the [5Xorb[105X package. You simply run the
  example in the following way:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/fi23m7/GOrbitByKOrbitsPrepare.g");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/fi23m7/GOrbitByKOrbitsSearch35.g");[127X[104X
    [4X[28X...[128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  will  not  go  into the details of the computation here, but they can be
  read off directly from the code in the files in that directory. In the first
  part,  run  by  the  file [11XGOrbitByKOrbitsPrepare.g[111X, we prepare the necessary
  input  data,  by  using similar techniques as described at length in Section
  [14X11.4[114X.  (Actually,  this example has been dealt with before the advent of the
  packages  [5Xchop[105X  and  [5Xgenss[105X,  hence  we  are  using  appropriate private code
  instead.)  We  are using two helper subgroups [22XU_1 < U_2 < H < Fi_{23}[122X, being
  [22X3[122X-subgroups  of  [22XH[122X  of order [22X81[122X and [22X6561[122X, respectively. The 1494-dimensional
  irreducible  representation  of  [22XFi_{23}[122X  over  the  field  with  2 elements
  contains  a  vector  that  is  fixed  by  [22XH[122X,  such  that  the  action on its
  [22XFi_{23}[122X-orbit is isomorphic to the action on the cosets of [22XH[122X.[133X
  
  [33X[0;0YThe  second  part,  in  the  file  [11XGOrbitByKOrbitsSearch35.g[111X,  is the actual
  enumeration of [22XH[122X-orbits:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[28Xsetup := OrbitBySuborbitBootstrapForVectors([128X[104X
    [4X[28X         [cu1gens,cu2gens,cngens],[u1gensp,u2gensp,ngensp],[128X[104X
    [4X[28X         [81,6561,3265173504],[10,30],rec());[128X[104X
    [4X[28Xobsol := InitOrbitBySuborbitList(setup,40);[128X[104X
    [4X[28Xl := Orb(cggens,v,OnRight,rec(schreier := true));[128X[104X
    [4X[28XEnumerate(l,100000);[128X[104X
    [4X[28XOrbitsFromSeedsToOrbitList(obsol,l);[128X[104X
    [4X[28Xorigseeds := List(obsol,OrigSeed); [128X[104X
    [4X[28Xpositions :=  List(origseeds,x->Position(l,x));[128X[104X
    [4X[28Xwords := List(positions,x->TraceSchreierTreeForward(l,x));[128X[104X
  [4X[32X[104X
  
  [33X[0;0YNote   that   this  computation  finds  only  35  of  the  36  double  coset
  representatives. The last corresponds to a very short suborbit which is very
  difficult  to  find.  Knowing  the  number  of  missing points, we guess the
  stabiliser in [22XH[122X of a missing representative, and find the latter amongst the
  fixed points of the stabiliser. We can then choose the one which lies in the
  [22XG[122X-orbit we have nearly enumerated above.[133X
  
  [33X[0;0YThese  double  coset  representatives were needed to determine the 2-modular
  character table of [22XFi_{23}[122X. Details of this can be found in [HNN06].[133X
  
  
  [1X11.3 [33X[0;0YThe Conway group [22XCo_1[122X[101X[1X acting in dimension [22X24[122X[101X[1X[133X[101X
  
  [33X[0;0YThe example in this section shows how to compute all suborbit lengths of the
  Conway  group  [22XCo_1[122X,  in  its  irreducible  projective action on a module of
  dimension  24 over the field with 5 elements. All the files for this example
  are  located  in the [11Xexamples/co1F5d24[111X subdirectory of the [5Xorb[105X package. Then
  you simply run the example in the following way:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/co1F5d24/Co1OrbitOnPF5d24.g");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/co1F5d24/Co1OrbitOnPF5d24.findall.g");[127X[104X
    [4X[28X...[128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  will  not go into the details of the first part of the computation here,
  as  they  are  very  similar to those reproduced in Section [14X11.1[114X, and can be
  directly read off from the code in the file [11XCo1OrbitOnPF5d24.g[111X: We are using
  three  helper  subgroups  [22XU_1  <  U_2  <  U_3  <  Co_1[122X, where [22XCo_1[122X has order
  [22X4157776806543360000≈  4.2⋅  10^18[122X,  see [CCN+85] or the [5XCTblLib[105X package, and
  where  [22XU_3≅  2_+^{1+8}.O_8(2)[122X  is the fifth maximal subgroup of [22XCo_1[122X, having
  order  [22X89181388800≈  8.9⋅  10^10[122X,  while  [22XU_2≅  [2^8]:  S_6(2)[122X  is a maximal
  subgroup  of  [22XU_3[122X  of  order [22X371589120≈ 3.7⋅ 10^8[122X, and [22XU_1≅ 2^6: L_3(2)[122X is a
  maximal  subgroup of [22XS_6(2)[122X of order [22X10752≈ 1.1⋅ 10^4[122X. The projective action
  comes from the irreducible 24-dimensional linear representation of the Schur
  cover   [22X2.Co_1[122X   of   [22XCo_1[122X,  which  by  [Jan05]  is  the  smallest  faithful
  representation  of  [22X2.Co_1[122X  over the field GF(5), and the quotient spaces we
  use for the helper subgroups have dimensions [22X8[122X, [22X8[122X and [22X16[122X respectively.[133X
  
  [33X[0;0YThe details of the second part can be directly read off from the code in the
  file [11XCo1OrbitOnPF5d24.findall.g[111X:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[28Xoo := InitOrbitBySuborbitList(setup,80);[128X[104X
    [4X[28Xl := MakeRandomLines(v,1000);[128X[104X
    [4X[28XOrbitsFromSeedsToOrbitList(oo,l);[128X[104X
    [4X[28Xintervecs := CVEC_ReadMatFromFile(Filename(DirectoriesPackageLibrary("orb",""),[128X[104X
    [4X[28X             "examples/co1F5d24/co1interestingvecs.cmat"));[128X[104X
    [4X[28XOrbitsFromSeedsToOrbitList(oo,intervecs);[128X[104X
    [4X[28XLength(oo!.obsos);[128X[104X
    [4X[28XSum(oo!.obsos,Size);[128X[104X
    [4X[28X(5^24-1)/(5-1);[128X[104X
  [4X[32X[104X
  
  [33X[0;0YNote that this example needs about 2GB of main memory on a 32bit machine and
  probably  nearly  4GB  on  a  64bit machine. However, the orbit lengths were
  previously  unknown  before  they were computed with this program. The orbit
  lengths were needed to rule out a case in [Mal06].[133X
  
  
  [1X11.4 [33X[0;0YThe Baby Monster [22XB[122X[101X[1X acting on its 2A involutions[133X[101X
  
  [33X[0;0YThe  example in this section shows how to enumerate the smallest non-trivial
  orbit  of  the  Baby  Monster  group  [22XB[122X.  All the files for this example are
  located  in  the [11Xexamples/bmF2d4370[111X subdirectory of the [5Xorb[105X package. You may
  simply run the example in the following way:[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/bmF2d4370/BMOrbitOnF2d4370partI.g");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XReadPackage("orb","examples/bmF2d4370/BMOrbitOnF2d4370partII.g");[127X[104X
    [4X[28X... [128X[104X
  [4X[32X[104X
  
  [33X[0;0YIn  the sequel we comment in detail on how the necessary input data actually
  is prepared. We begin by loading the packages we are going to use.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XLoadPackage("orb");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XLoadPackage("io");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XLoadPackage("cvec");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XLoadPackage("atlasrep");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XLoadPackage("chop");[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XLoadPackage("genss");[127X[104X
    [4X[28X...  [128X[104X
  [4X[32X[104X
  
  [33X[0;0YThe  one-point stabilisers associated to the smallest non-trivial orbit of [22XB[122X
  are   its  largest  maximal  subgroups  [22XE  ≅  2.^2E_6(2).2[122X,  which  are  the
  centralisers  of  its  2A involutions. Here [22XE[122X is a bicyclic extension of the
  twisted  Lie  type  group [22X^2E_6(2)[122X, and has index [22X[B: E]=13571955000 ≈ 1.4 ⋅
  10^10[122X, see [CCN+85] or the [5XCTblLib[105X package.[133X
  
  [33X[0;0YWe  first  try to find a matrix representation of [22XB[122X such that the [22XB[122X-orbit we
  look for is realised as a set of vectors in the underlying vector space. The
  smallest  faithful  representation  of  [22XB[122X  over  the field GF(2), by [Jan05]
  having  dimension  4370,  springs  to  mind.  Explicit  matrices in terms of
  standard  generators in the sense of [Wil96] are available in [Wil], and are
  accessibe through the [5Xatlasrep[105X package. Moreover, we find generators of [22XE[122X by
  applying  a  straight  line program, also available in the [5Xatlasrep[105X package,
  expressing generators of [22XE[122X in terms of the generators of [22XB[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xgens := AtlasGenerators([ "B", [ "BG1-f2r4370B0.m1", "BG1-f2r4370B0.m2" ], 1, 2 ]).generators;[127X[104X
    [4X[28X[ <an immutable 4370x4370 matrix over GF2>, [128X[104X
    [4X[28X  <an immutable 4370x4370 matrix over GF2> ][128X[104X
    [4X[25Xgap>[125X [27Xbgens := List(gens,CMat);[127X[104X
    [4X[28X[ <cmat 4370x4370 over GF(2,1)>, <cmat 4370x4370 over GF(2,1)> ] [128X[104X
    [4X[25Xgap>[125X [27Xslpbtoe := AtlasStraightLineProgram([ "B", "BG1-max1W1", 1 ]).program;;[127X[104X
    [4X[25Xgap>[125X [27Xegens := ResultOfStraightLineProgram(slpbtoe,bgens);[127X[104X
    [4X[28X[ <cmat 4370x4370 over GF(2,1)>, <cmat 4370x4370 over GF(2,1)> ] [128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  look for a non-zero vector being fixed by both generators of [22XE[122X. It turns
  out  that the latter have a common fixed space of dimension 1. Then, since [22XE[122X
  is  a maximal subgroup, the stabiliser in [22XB[122X of the non-zero vector [22Xv[122X in that
  fixed space coincides with [22XE[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xx := egens[1]-egens[1]^0;;[127X[104X
    [4X[25Xgap>[125X [27Xnsx := NullspaceMat(x);[127X[104X
    [4X[28X<immutable cmat 2202x4370 over GF(2,1)>[128X[104X
    [4X[25Xgap>[125X [27Xy := nsx * (egens[2]-egens[2]^0);;[127X[104X
    [4X[25Xgap>[125X [27Xnsy := NullspaceMat(y);[127X[104X
    [4X[28X<immutable cmat 1x2202 over GF(2,1)>[128X[104X
    [4X[25Xgap>[125X [27Xv := nsy[1]*nsx;[127X[104X
    [4X[28X<immutable cvec over GF(2,1) of length 4370> [128X[104X
  [4X[32X[104X
  
  [33X[0;0YStoring  eight  elements  of  GF(2) into 1 byte, to store a vector of length
  4370  needs  547  bytes plus some organisational overhead resulting in about
  580 bytes, hence to store the full [22XB[122X-orbit of [22Xv[122X we need [22X580 ⋅ [B: E] ≈ 7.9 ⋅
  10^12[122X  bytes.  Hence  we  try to find helper subgroups suitable to achieve a
  saving  factor  of  [22X≈  10^4[122X,  i. e. allowing to store only one out of [22X≈ 10^4[122X
  vectors.  To  this  end, we look for a pair [22XU_1<U_2[122X of helper subgroups such
  that  [22X|U_2|  ≈  10^5[122X,  where we take into account that typically the overall
  saving  factor  achieved  is  somewhat smaller than the order of the largest
  helper subgroup.[133X
  
  [33X[0;0YBy  [CCN+85], and a few computations with subgroup fusions using the [5XCTblLib[105X
  package,  the  derived  subgroup  [22XE'  ≅ 2.^2E_6(2)[122X of [22XE[122X turns out to possess
  maximal  subgroups  [22X2  × Fi_{22}[122X and [22X2.Fi_{22}[122X, where [22XFi_{22}[122X denotes one of
  the sporadic simple Fischer groups, and where the former constitute a unique
  conjugacy  class  with  associated  normalizers in [22XE[122X of shape [22X2 × Fi_{22}.2[122X,
  while the latter consist of two conjugacy classes being self-normalising and
  interchanged by [22XE[122X.[133X
  
  [33X[0;0YNow  [22XFi_{22}[122X has a unique conjugacy class of maximal subgroups [22XM_{12}[122X, where
  the  latter denotes one of the sporadic simple Mathieu groups; the subgroups
  [22XM_{12}[122X  lift  to  a unique conjugacy class of subgroups [22XM_{12}[122X of [22X2.Fi_{22}[122X,
  which  turn  out to constitute a conjugacy class of subgroups of [22XE[122X different
  from  the  subgroups  [22XM_{12}[122X  being  contained  in  [22XFi_{22}[122X. Anyway, we have
  [22X|M_{12}|=95040[122X, hence [22XU_2=M_{12}[122X seems to be a good candidate for the larger
  helper subgroup. In particular, there is a unique conjugacy class of maximal
  subgroups   [22XL_2(11)[122X   of   [22XM_{12}[122X,  and  since  [22X|L_2(11)|=660[122X  and  [22X[M_{12}:
  L_2(11)]=144[122X  letting  [22XU_1=L_2(11)[122X  seems  to  be  a  good candidate for the
  smaller helper subgroup. Recall that [22XU_1[122X and [22XU_2[122X are useful helper subgroups
  only  if  we  are  able  to  find suitable quotient modules allowing for the
  envisaged saving factor.[133X
  
  [33X[0;0YTo find [22XU_1[122X and [22XU_2[122X, we first try to find a subgroup [22XFi_{22}[122X or [22X2.Fi_{22}[122X of
  [22XE[122X. We start a random search, aiming at finding standard generators of either
  [22XFi_{22}[122X or [22X2.Fi_{22}[122X, and we use [10XGeneratorsWithMemory[110X in order to be able to
  express  the generators found as words in the generators of [22XE[122X. To accelerate
  computations  we first construct a small representation of [22XE[122X; by [Jan05] the
  smallest  faithful  irreducible  representation  of  [22XFi_{22}[122X  over GF(2) has
  dimension  78,  hence  we  cannot do better for [22XE[122X; note that the latter is a
  representation of [22XoverlineE:=E/Z(E) ≅ ^2E_6(2).2[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XSetInfoLevel(InfoChop,2);[127X[104X
    [4X[25Xgap>[125X [27Xm := Module(egens);[127X[104X
    [4X[28X<module of dim. 4370 over GF(2)>[128X[104X
    [4X[25Xgap>[125X [27Xr := Chop(m);[127X[104X
    [4X[28X...[128X[104X
    [4X[28Xrec( ischoprecord := true, [128X[104X
    [4X[28X  db := [ <abs. simple module of dim. 78 over GF(2)>, [128X[104X
    [4X[28X          <trivial module of dim. 1 over GF(2)>, [128X[104X
    [4X[28X          <abs. simple module of dim. 1702 over GF(2)>, [128X[104X
    [4X[28X          <abs. simple module of dim. 572 over GF(2)> ], [128X[104X
    [4X[28X  mult := [ 5, 4, 2, 1 ], acs := [ 1, 2, 3, 1, 4, 1, 1, 2, 2, 3, 1, 2 ], [128X[104X
    [4X[28X  module := <reducible module of dim. 4370 over GF(2)> )[128X[104X
    [4X[25Xgap>[125X [27Xi := Position(List(r.db,Dimension),78);;[127X[104X
    [4X[25Xgap>[125X [27Xegens78 := GeneratorsWithMemory(RepresentingMatrices(r.db[i]));[127X[104X
    [4X[28X[ <<immutable cmat 78x78 over GF(2,1)> with mem>, [128X[104X
    [4X[28X  <<immutable cmat 78x78 over GF(2,1)> with mem> ] [128X[104X
  [4X[32X[104X
  
  [33X[0;0YBy  [Wil],  standard generators [22Xa,b[122X of [22XFi_{22}[122X are given as follows: [22Xa[122X is an
  element   of   the   2A   conjugacy   class  of  [22XFi_{22}[122X,  and  [22Xb[122X,  [22Xab[122X,  and
  [22X(ab)^4bab(abb)^2[122X   have   order  13,  11,  and  12,  respectively;  standard
  generators  of  [22X2.Fi_{22}[122X are lifts of standard generators of [22XFi_{22}[122X having
  the same order fingerprint. The 2A conjugacy class of [22XFi_{22}[122X fuses into the
  2A  conjugacy  class of [22XoverlineE[122X, where the latter is obtained as the 11-th
  power  of  the unique conjugacy class of elements of order 22, and [22XoverlineE[122X
  has only one conjugacy class of elements of order 13.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xo := Orb(egens78,StripMemory(egens78[1])^0,OnRight,rec(schreier := true,[127X[104X
    [4X[25X>[125X [27X            lookingfor := function(o,x) return Order(x)=22; end));[127X[104X
    [4X[28X<open orbit, 1 points with Schreier tree looking for sth.>[128X[104X
    [4X[25Xgap>[125X [27XEnumerate(o);[127X[104X
    [4X[28X<open orbit, 393 points with Schreier tree looking for sth.>[128X[104X
    [4X[25Xgap>[125X [27Xword := TraceSchreierTreeForward(o,PositionOfFound(o));[127X[104X
    [4X[28X[ 1, 2, 1, 2, 2, 1, 2, 2, 1, 2, 2 ][128X[104X
    [4X[25Xgap>[125X [27Xg2a := Product(egens78{word})^11;[127X[104X
    [4X[28X<<immutable cmat 78x78 over GF(2,1)> with mem>[128X[104X
    [4X[25Xgap>[125X [27Xo := Orb(egens78,StripMemory(egens78[1])^0,OnRight,rec(schreier := true,[127X[104X
    [4X[25X>[125X [27X            lookingfor := function(o,x) return Order(x)=13; end));[127X[104X
    [4X[28X<open orbit, 1 points with Schreier tree looking for sth.>[128X[104X
    [4X[25Xgap>[125X [27XEnumerate(o);[127X[104X
    [4X[28X<open orbit, 144 points with Schreier tree looking for sth.>[128X[104X
    [4X[25Xgap>[125X [27Xword := TraceSchreierTreeForward(o,PositionOfFound(o));[127X[104X
    [4X[28X[ 1, 2, 1, 2, 1, 2, 1, 2, 2 ][128X[104X
    [4X[25Xgap>[125X [27Xb := Product(egens78{word});[127X[104X
    [4X[28X<<immutable cmat 78x78 over GF(2,1)> with mem> [128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  search through the [22XoverlineE[122X-conjugates of [10Xg2a[110X until we find a conjugate
  [22Xa[122X  together with [22Xb[122X fulfilling the defining conditions of standard generators
  of  [22XFi_{22}[122X,  and  moreover  fulfilling  the  relations  of  the  associated
  presentation of [22XFi_{22}[122X available in [Wil].[133X
  
  [33X[0;0YTo  find  conjugates,  we  use  the product replacement algorithm to produce
  pseudo  random  elements of [22XoverlineE[122X. Assuming a genuine random search, the
  success    probability   of   this   approach   is   as   follows:   Letting
  [22XoverlineE':=E'/Z(E')        ≅        ^2E_6(2)[122X,        out       of       the
  [22X|overlineE'|/|C_overlineE'}(g2a)|[122X   conjugates   of   [10Xg2a[110X   there  are  [22X|C_{
  overlineE' }(b)|/|C_{ overlineE' }(Fi_{22})| =|C_{ overlineE' }(b)|[122X elements
  together  with  the  fixed  element [22Xb[122X giving standard generators of [22XFi_{22}[122X.
  Since  [22XFi_{22}[122X  has two conjugacy classes of elements of order 13, and there
  are  three conjugacy classes of subgroups [22XFi_{22}[122X of [22XoverlineE'[122X, the success
  probability   is   [22X6   ⋅   |C_{   overlineE'   }(g2a)|   ⋅  |C_{  overlineE'
  }(b)|/|overlineE'| ≈ 2 ⋅ 10^-5[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xpr := ProductReplacer(egens78,rec(maxdepth := 150));[127X[104X
    [4X[28X<product replacer nrgens=2 slots=12 scramble=100 maxdepth=150 steps=0 (rattle)>[128X[104X
    [4X[25Xgap>[125X [27Xi := 0;;[127X[104X
    [4X[25Xgap>[125X [27Xrepeat[127X[104X
    [4X[25X>[125X [27X     i := i + 1; [127X[104X
    [4X[25X>[125X [27X     x := Next(pr);[127X[104X
    [4X[25X>[125X [27X     a := g2a^x;[127X[104X
    [4X[25X>[125X [27X   until IsOne((a*b)^11) and IsOne(((a*b)^4*b*a*b*(a*b*b)^2)^12) and[127X[104X
    [4X[25X>[125X [27X     IsOne((a*b^2)^21) and IsOne(Comm(a,b)^3) and [127X[104X
    [4X[25X>[125X [27X     IsOne(Comm(a,b^2)^3) and IsOne(Comm(a,b^3)^3) and[127X[104X
    [4X[25X>[125X [27X     IsOne(Comm(a,b^4)^2) and IsOne(Comm(a,b^5)^3) and[127X[104X
    [4X[25X>[125X [27X     IsOne(Comm(a,b*a*b^2)^3) and IsOne(Comm(a,b^-1*a*b^-2)^2) and[127X[104X
    [4X[25X>[125X [27X     IsOne(Comm(a,b*a*b^5)^2) and IsOne(Comm(a,b^2*a*b^5)^2);[127X[104X
    [4X[25Xgap>[125X [27Xi;[127X[104X
    [4X[28X53271 [128X[104X
  [4X[32X[104X
  
  [33X[0;0YNote  that  the  initial state of the random number generator does influence
  this  randomised  result:  it may very well be that you see some other value
  for [22Xi[122X.[133X
  
  [33X[0;0YDue to a presentation being available we have proved that the elements found
  generate  a  subgroup  [22XFi_{22}[122X. If we had not had a presentation at hand, we
  might  only  have  been  able  to  find  elements  fulfilling  the  defining
  conditions  of  standard  generators  of  [22XFi_{22}[122X,  but  still  generating a
  subgroup  of  another  isomorphism type. In that case, for further checks we
  can  use  the  following tools: We try to find a short orbit of vectors, and
  using a randomized Schreier-Sims algorithm gives a lower bound for the order
  of  the  group  seen.  However,  we can use the action on the orbit to get a
  homomorphism  into  a  permutation  group,  allowing to prove that the group
  generated indeed has [22XFi_{22}[122X as a quotient.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XS := StabilizerChain(Group(a,b),rec(TryShortOrbit := 30,[127X[104X
    [4X[25X>[125X [27X                                       OrbitLengthLimit := 5000));[127X[104X
    [4X[28X...[128X[104X
    [4X[28X<stabchain size=64561751654400 orblen=3510 layer=1 SchreierDepth=8>[128X[104X
    [4X[28X <stabchain size=18393661440 orblen=2816 layer=2 SchreierDepth=7>[128X[104X
    [4X[28X  <stabchain size=6531840 orblen=1680 layer=3 SchreierDepth=7>[128X[104X
    [4X[28X   <stabchain size=3888 orblen=243 layer=4 SchreierDepth=5>[128X[104X
    [4X[28X    <stabchain size=16 orblen=16 layer=5 SchreierDepth=2>[128X[104X
    [4X[25Xgap>[125X [27XSize(S)=Size(CharacterTable("Fi22"));[127X[104X
    [4X[28Xtrue [128X[104X
    [4X[25Xgap>[125X [27Xp := Group(ActionOnOrbit(S!.orb,[a,b]));;[127X[104X
    [4X[25Xgap>[125X [27XDisplayCompositionSeries(p);[127X[104X
    [4X[28XG (2 gens, size 64561751654400)[128X[104X
    [4X[28X | Fi(22)[128X[104X
    [4X[28X1 (0 gens, size 1) [128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe now return to our original representation.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XSetInfoLevel(InfoSLP,2); [127X[104X
    [4X[25Xgap>[125X [27Xslpetofi22 := SLPOfElms([a,b]);[127X[104X
    [4X[28X<straight line program>[128X[104X
    [4X[25Xgap>[125X [27XLength(LinesOfStraightLineProgram(slpetofi22));[127X[104X
    [4X[28X278[128X[104X
    [4X[25Xgap>[125X [27XSlotUsagePattern(slpetofi22);;[127X[104X
    [4X[25Xgap>[125X [27Xfgens := ResultOfStraightLineProgram(slpetofi22,egens);[127X[104X
    [4X[28X...[128X[104X
    [4X[28X[ <cmat 4370x4370 over GF(2,1)>, <cmat 4370x4370 over GF(2,1)> ] [128X[104X
    [4X[25Xgap>[125X [27Xa := fgens[1];;[127X[104X
    [4X[25Xgap>[125X [27Xb := fgens[2];; [127X[104X
    [4X[25Xgap>[125X [27XIsOne(b^13);[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[25Xgap>[125X [27XIsOne((a*b)^11);[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[25Xgap>[125X [27XIsOne((a*b^2)^21);[127X[104X
    [4X[28Xtrue [128X[104X
  [4X[32X[104X
  
  [33X[0;0YBy  construction  the  group  generated  by [22Xa,b[122X is [22XFi_{22}[122X or [22X2 × Fi_{22}[122X or
  [22X2.Fi_{22}[122X.  Note  that due to different seeds in the random number generator
  it is in fact possible at this stage that you have created a different group
  as displayed here! In our outcome, since [22Xa[122X has even order, and both [22Xb[122X and [22Xab[122X
  have odd order, we cannot possibly have [22X2 × Fi_{22}[122X; and by the presentation
  of  [22X2.Fi_{22}[122X  available in [Wil] the order of [22Xab^2[122X being 21 implies that we
  cannot  possibly  have [22X2.Fi_{22}[122X either. Hence we indeed have found standard
  generators  of  [22XFi_{22}[122X.  If  we  had  hit  one  of the cases [22X2 × Fi_{22}[122X or
  [22X2.Fi_{22}[122X,  we could just continue the above search until we find a subgroup
  [22XFi_{22}[122X,  or  using  the  above order fingerprint we could easily modify the
  elements found to obtain standard generators of either [22XFi_{22}[122X or [22X2.Fi_{22}[122X.[133X
  
  [33X[0;0YNow,  standard  generators  of [22XU_2=M_{12}[122X in terms of standard generators of
  [22XFi_{22}[122X,  and  generators  of [22XU_1=L_2(11)[122X in terms of standard generators of
  [22XM_{12}[122X  are  accessible in the [5Xatlasrep[105X package. Note that if we had found a
  subgroup  [22X2.Fi_{22}[122X  above,  since  [22XM_{12}[122X lifts to a subgroup [22X2 × M_{12}[122X of
  [22X2.Fi_{22}[122X, it would again be easy to find standard generators of [22XM_{12}[122X from
  the generators of [22XM_{12}[122X or [22X2 × M_{12}[122X respectively provided by the [5Xatlasrep[105X
  package.  Anyway,  the  next task is to find good quotient modules such that
  the  helper  subgroups  have  longish  orbits  on  vectors.  To this end, we
  restrict to [22XM_{12}[122X and compute the radical series of the restricted module.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xslpfi22tom12 := AtlasStraightLineProgram([ "Fi22", "F22G1-max14W1", 1 ]).program;;[127X[104X
    [4X[25Xgap>[125X [27Xslpm12tol211 := AtlasStraightLineProgram([ "M12", "M12G1-max5W1", 1 ]).program;;[127X[104X
    [4X[25Xgap>[125X [27Xmgens := ResultOfStraightLineProgram(slpfi22tom12,fgens);[127X[104X
    [4X[28X[ <cmat 4370x4370 over GF(2,1)>, <cmat 4370x4370 over GF(2,1)> ][128X[104X
    [4X[25Xgap>[125X [27Xlgens := ResultOfStraightLineProgram(slpm12tol211,mgens);[127X[104X
    [4X[28X[ <cmat 4370x4370 over GF(2,1)>, <cmat 4370x4370 over GF(2,1)> ] [128X[104X
    [4X[25Xgap>[125X [27Xm := Module(mgens);;[127X[104X
    [4X[25Xgap>[125X [27Xr := Chop(m);;[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27Xrad := RadicalSeries(m,r.db);[127X[104X
    [4X[28X...[128X[104X
    [4X[28Xrec( [128X[104X
    [4X[28X  db := [ <abs. simple module of dim. 144 over GF(2)>, [128X[104X
    [4X[28X          <abs. simple module of dim. 44 over GF(2)>, [128X[104X
    [4X[28X          <simple module of dim. 32 over GF(2) splitting field degree 2>, [128X[104X
    [4X[28X          <abs. simple module of dim. 10 over GF(2)>, [128X[104X
    [4X[28X          <trivial module of dim. 1 over GF(2)> ],[128X[104X
    [4X[28X  module := <reducible module of dim. 4370 over GF(2)>,[128X[104X
    [4X[28X  basis := <immutable cmat 4370x4370 over GF(2,1)>,[128X[104X
    [4X[28X  ibasis := <immutable cmat 4370x4370 over GF(2,1)>,[128X[104X
    [4X[28X  cfposs := [ [ [ 1 .. 144 ], [ 145 .. 288 ], [ 289 .. 432 ], [ 433 .. 576 ],[128X[104X
    [4X[28X...[128X[104X
    [4X[28X  isotypes := [ [ 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3,[128X[104X
    [4X[28X                  3, 3, 3, 3, 3, 3, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5 ],[128X[104X
    [4X[28X...[128X[104X
    [4X[28X  isradicalrecord := true ) [128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  observe  that there are faithful irreducible quotients of dimensions 10,
  32,  44,  and  144. Since we look for a quotient module such that [22XM_{12}[122X has
  many  regular  orbits  on  vectors,  we  ignore  the  irreducible  module of
  dimension 10. We consider the one of dimension 32.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xi := Position(List(rad.db,Dimension),32);;[127X[104X
    [4X[25Xgap>[125X [27Xmgens32 := RepresentingMatrices(rad.db[i]);[127X[104X
    [4X[28X[ <immutable cmat 32x32 over GF(2,1)>, <immutable cmat 32x32 over GF(2,1)> ][128X[104X
    [4X[25Xgap>[125X [27XOrbitStatisticOnVectorSpace(mgens32,95040,30);[127X[104X
    [4X[28XFound length     95040, have now   24 orbits, average length: 93060 [128X[104X
  [4X[32X[104X
  
  [33X[0;0YThis  is  excellent  indeed.  Hence we pick a summand of dimension 32 in the
  first  radical  layer,  and  apply  the  associated  base  change to all the
  generators.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xbgens := List(bgens,x->rad.basis*x*rad.ibasis);;[127X[104X
    [4X[25Xgap>[125X [27Xegens := List(egens,x->rad.basis*x*rad.ibasis);;[127X[104X
    [4X[25Xgap>[125X [27Xfgens := List(fgens,x->rad.basis*x*rad.ibasis);;[127X[104X
    [4X[25Xgap>[125X [27Xmgens := List(mgens,x->rad.basis*x*rad.ibasis);;[127X[104X
    [4X[25Xgap>[125X [27Xlgens := List(lgens,x->rad.basis*x*rad.ibasis);;[127X[104X
    [4X[25Xgap>[125X [27Xj := Position(rad.isotypes[1],i);;[127X[104X
    [4X[25Xgap>[125X [27Xl := rad.cfposs[1][j];;[127X[104X
    [4X[25Xgap>[125X [27XAppend(l,Difference([1..4370],l));[127X[104X
    [4X[25Xgap>[125X [27Xbgens := List(bgens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xegens := List(egens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xfgens := List(fgens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xmgens := List(mgens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xlgens := List(lgens,x->ORB_PermuteBasisVectors(x,l));; [127X[104X
  [4X[32X[104X
  
  [33X[0;0YWe consider the irreducible quotient module of [22XM_{12}[122X of dimension 32, whose
  restriction  to  [22XL_2(11)[122X  turns  out  to  be  is semisimple. The irreducible
  quotients of dimension 10 are too small to have too many regular orbits, but
  the direct sum of two of them turns out to work fine.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xlgens32 := List(lgens,x->ExtractSubMatrix(x,[1..32],[1..32]));[127X[104X
    [4X[28X[ <cmat 32x32 over GF(2,1)>, <cmat 32x32 over GF(2,1)> ][128X[104X
    [4X[25Xgap>[125X [27Xm := Module(lgens32);;[127X[104X
    [4X[25Xgap>[125X [27Xr := Chop(m);[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27Xsoc := SocleSeries(m,r.db);[127X[104X
    [4X[28X...[128X[104X
    [4X[28Xrec( issoclerecord := true,[128X[104X
    [4X[28X  db := [ <simple module of dim. 10 over GF(2) splitting field degree 2>,[128X[104X
    [4X[28X          <trivial module of dim. 1 over GF(2)>,[128X[104X
    [4X[28X          <abs. simple module of dim. 10 over GF(2)> ],[128X[104X
    [4X[28X  module := <reducible module of dim. 32 over GF(2)>,[128X[104X
    [4X[28X  basis := <cmat 32x32 over GF(2,1)>, ibasis := <cmat 32x32 over GF(2,1)>,[128X[104X
    [4X[28X  cfposs := [ [ [ 1 .. 10 ], [ 11 ], [ 12 ], [ 13 .. 22 ], [ 23 .. 32 ] ] ],[128X[104X
    [4X[28X  isotypes := [ [ 1, 2, 2, 3, 3 ] ] )[128X[104X
    [4X[25Xgap>[125X [27Xi := Position(List(soc.db,x->[Dimension(x),DegreeOfSplittingField(x)]),[127X[104X
    [4X[25X>[125X [27X                                [10,1]);;[127X[104X
    [4X[25Xgap>[125X [27Xj := Position(soc.isotypes[1],i);;[127X[104X
    [4X[25Xgap>[125X [27Xl := Concatenation(soc.cfposs[1]{[j,j+1]});;[127X[104X
    [4X[25Xgap>[125X [27Xlgens32 := List(lgens32,x->soc.basis*x*soc.ibasis);[127X[104X
    [4X[28X[ <cmat 32x32 over GF(2,1)>, <cmat 32x32 over GF(2,1)> ][128X[104X
    [4X[25Xgap>[125X [27Xlgens20 := List(lgens32,x->ExtractSubMatrix(x,l,l));[127X[104X
    [4X[28X[ <cmat 20x20 over GF(2,1)>, <cmat 20x20 over GF(2,1)> ][128X[104X
    [4X[25Xgap>[125X [27XOrbitStatisticOnVectorSpace(lgens20,660,30);[127X[104X
    [4X[28XFound length       660, have now 4401 orbits, average length: 598 [128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe apply the appropriate base change to all the generators.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xt := ORB_EmbedBaseChangeTopLeft(soc.basis,4370);[127X[104X
    [4X[28X<cmat 4370x4370 over GF(2,1)>[128X[104X
    [4X[25Xgap>[125X [27Xti := ORB_EmbedBaseChangeTopLeft(soc.ibasis,4370);[127X[104X
    [4X[28X<cmat 4370x4370 over GF(2,1)>[128X[104X
    [4X[25Xgap>[125X [27Xbgens := List(bgens,x->t*x*ti);;[127X[104X
    [4X[25Xgap>[125X [27Xegens := List(egens,x->t*x*ti);;[127X[104X
    [4X[25Xgap>[125X [27Xfgens := List(fgens,x->t*x*ti);;[127X[104X
    [4X[25Xgap>[125X [27Xmgens := List(mgens,x->t*x*ti);;[127X[104X
    [4X[25Xgap>[125X [27Xlgens := List(lgens,x->t*x*ti);; [127X[104X
    [4X[25Xgap>[125X [27XAppend(l,Difference([1..4370],l));[127X[104X
    [4X[25Xgap>[125X [27Xbgens := List(bgens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xegens := List(egens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xfgens := List(fgens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xmgens := List(mgens,x->ORB_PermuteBasisVectors(x,l));;[127X[104X
    [4X[25Xgap>[125X [27Xlgens := List(lgens,x->ORB_PermuteBasisVectors(x,l));; [127X[104X
  [4X[32X[104X
  
  [33X[0;0YHaving reached the ultimate choice of basis, we recreate the fixed vector [10Xv[110X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xx := egens[1]-egens[1]^0;;[127X[104X
    [4X[25Xgap>[125X [27Xnsx := NullspaceMat(x);;[127X[104X
    [4X[25Xgap>[125X [27Xy := nsx * (egens[2]-egens[2]^0);;[127X[104X
    [4X[25Xgap>[125X [27Xnsy := NullspaceMat(y);;[127X[104X
    [4X[25Xgap>[125X [27Xv := nsy[1]*nsx;; [127X[104X
  [4X[32X[104X
  
  [33X[0;0YFinally  we  need  small  faithful permutation representations of the helper
  subgroups.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xmgens32 := List(mgens,x->ExtractSubMatrix(x,[1..32],[1..32]));[127X[104X
    [4X[28X[ <cmat 32x32 over GF(2,1)>, <cmat 32x32 over GF(2,1)> ][128X[104X
    [4X[25Xgap>[125X [27XS := StabilizerChain(Group(mgens32),rec(TryShortOrbit := 10));[127X[104X
    [4X[28X...[128X[104X
    [4X[28X<stabchain size=95040 orblen=3960 layer=1 SchreierDepth=7>[128X[104X
    [4X[28X <stabchain size=24 orblen=24 layer=2 SchreierDepth=4>[128X[104X
    [4X[25Xgap>[125X [27Xp := Group(ActionOnOrbit(S!.orb,mgens32));[127X[104X
    [4X[28X<permutation group with 2 generators>[128X[104X
    [4X[25Xgap>[125X [27Xi := SmallerDegreePermutationRepresentation(p);;[127X[104X
    [4X[25Xgap>[125X [27Xpp := Group(List(GeneratorsOfGroup(p),x->ImageElm(i,x)));[127X[104X
    [4X[28X<permutation group with 2 generators>[128X[104X
    [4X[25Xgap>[125X [27Xm12 := MathieuGroup(12);;[127X[104X
    [4X[25Xgap>[125X [27Xi := IsomorphismGroups(pp,m12);;[127X[104X
    [4X[25Xgap>[125X [27Xmpermgens := List(GeneratorsOfGroup(pp),x->ImageElm(i,x));[127X[104X
    [4X[28X[ (5,7)(6,11)(8,9)(10,12), (1,10,3)(2,11,12)(4,5,6)(7,9,8) ][128X[104X
    [4X[25Xgap>[125X [27Xlpermgens := ResultOfStraightLineProgram(slpm12tol211,mpermgens);[127X[104X
    [4X[28X[ (1,8)(2,5)(3,9)(4,7)(6,11)(10,12), (1,8,3)(2,7,12)(4,6,9)(5,11,10) ] [128X[104X
  [4X[32X[104X
  
  [33X[0;0YWe  could  just go on from here, however, sometimes it is useful to save all
  the created data to disk.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xf := IO_File("data.gp","w");;[127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,"seed");; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,v);; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,"generators");; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,bgens);; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,egens);; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,fgens);; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,mgens);; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,lgens);;[127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,"permutations");; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,mpermgens);; [127X[104X
    [4X[25Xgap>[125X [27XIO_Pickle(f,lpermgens);;[127X[104X
    [4X[25Xgap>[125X [27XIO_Close(f);; [127X[104X
  [4X[32X[104X
  
  [33X[0;0YThis can be loaded again, in particular into a new [5XGAP[105X session, as follows.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XLoadPackage("orb");;[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27XLoadPackage("cvec");;[127X[104X
    [4X[28X...[128X[104X
    [4X[25Xgap>[125X [27Xf := IO_File("data.gp");[127X[104X
    [4X[28X<file fd=4 rbufsize=65536 rpos=1 rdata=0>[128X[104X
    [4X[25Xgap>[125X [27XIO_Unpickle(f);[127X[104X
    [4X[28X"seed"[128X[104X
    [4X[25Xgap>[125X [27Xv:=IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27XIO_Unpickle(f);[127X[104X
    [4X[28X"generators"[128X[104X
    [4X[25Xgap>[125X [27Xbgens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27Xegens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27Xfgens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27Xmgens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27Xlgens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27XIO_Unpickle(f);[127X[104X
    [4X[28X"permutations"[128X[104X
    [4X[25Xgap>[125X [27Xmpermgens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27Xlpermgens := IO_Unpickle(f);;[127X[104X
    [4X[25Xgap>[125X [27XIO_Close(f);; [127X[104X
  [4X[32X[104X
  
  [33X[0;0YNow we are prepared to actually run the orbit enumeration. Note that for the
  following  memory  estimates we assume that we are running things on a 64bit
  machine.  On a 32bit machine the overhead is smaller. We expect that all the
  vectors  in  the smaller quotient of dimension 20 will enumerated; needing 3
  bytes  per  vector  for  the actual data which results in 40 bytes including
  overhead,  this  amounts  to [22X40 ⋅ 2^20 ≈ 42[122X MB of memory space. Since [22X2^32 ≈
  4.3  ⋅  10^9[122X is less than [22X[B: E][122X, we also expect that the larger quotient of
  dimension  32  will  be  enumerated completely, by [22XL_2(11)[122X-orbits; needing 4
  bytes  per  vector  for  the  actual  data  resulting  in 40 bytes including
  overhead,    and    assuming    a    saving    factor    as   suggested   by
  [10XOrbitStatisticOnVectorSpace[110X  yields  an estimated memory requirement of [22X40 ⋅
  2^32  ⋅  1/598  ≈  287[122X  MB.  For  the  large  [22XB[122X-orbit,  being  enumerated by
  [22XM_{12}[122X-orbits, we similarly get an estimated memory requirement of [22X584 ⋅ [B:
  E] ⋅ 1/93060 ≈ 85[122X MB.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xsetup := OrbitBySuborbitBootstrapForVectors([127X[104X
    [4X[25X>[125X [27X            [lgens,mgens,bgens],[lpermgens,mpermgens,[(),()]],[127X[104X
    [4X[25X>[125X [27X            [660,95040,4154781481226426191177580544000000],[20,32],rec());[127X[104X
    [4X[28X#I  Calculating stabilizer chain for whole group...[128X[104X
    [4X[28X#I  Trying smaller degree permutation representation for U2...[128X[104X
    [4X[28X#I  Trying smaller degree permutation representation for U1...[128X[104X
    [4X[28X#I  Enumerating permutation base images of U_1...[128X[104X
    [4X[28X#I  Looking for U1-coset-recognising U2-orbit in factor space...[128X[104X
    [4X[28X#I  OrbitBySuborbit found 100% of a U2-orbit of size 95 040[128X[104X
    [4X[28X#I  Found 144 suborbits (need 144)[128X[104X
    [4X[28X<setup for an orbit-by-suborbit enumeration, k=2>[128X[104X
    [4X[25Xgap>[125X [27Xo := OrbitBySuborbitKnownSize(setup,v,3,3,2,51,13571955000);[127X[104X
    [4X[28X#I  OrbitBySuborbit found 100% of a U2-orbit of size 1[128X[104X
    [4X[28X#I  OrbitBySuborbit found 100% of a U2-orbit of size 23 760[128X[104X
    [4X[28X...[128X[104X
    [4X[28X#I  OrbitBySuborbit found 51% of a U3-orbit of size 13 571 955 000[128X[104X
    [4X[28X<orbit-by-suborbit size=13571955000 stabsize=306129918735099415756800 ([128X[104X
    [4X[28X51%) saving factor=56404> [128X[104X
  [4X[32X[104X
  
  [33X[0;0YIndeed the saving factor actually achieved is smaller than the best possible
  estimate given above, but it still has the same order of magnitude.[133X
  
