  
  [1X18 [33X[0;0YCyclotomic Numbers[133X[101X
  
  [33X[0;0Y[5XGAP[105X  admits  computations in abelian extension fields of the rational number
  field  [22Xℚ[122X,  that is fields with abelian Galois group over [22Xℚ[122X. These fields are
  subfields  of  [13Xcyclotomic  fields[113X  [22Xℚ(e_n)[122X  where  [22Xe_n  =  exp(2  π i/n)[122X is a
  primitive  complex  [22Xn[122X-th  root  of  unity.  The elements of these fields are
  called [13Xcyclotomics[113X.[133X
  
  [33X[0;0YInformation  concerning  operations  for domains of cyclotomics, for example
  certain integral bases of fields of cyclotomics, can be found in Chapter [14X60[114X.
  For  more  general  operations  that  take  a field extension as a –possibly
  optional–  argument,  e.g.,  [2XTrace[102X  ([14X58.3-5[114X)  or  [2XCoefficients[102X ([14X61.6-3[114X), see
  Chapter [14X58[114X.[133X
  
  
  [1X18.1 [33X[0;0YOperations for Cyclotomics[133X[101X
  
  [1X18.1-1 E[101X
  
  [29X[2XE[102X( [3Xn[103X ) [32X operation
  
  [33X[0;0Y[2XE[102X  returns  the  primitive [3Xn[103X-th root of unity [22Xe_n = exp(2π i/n)[122X. Cyclotomics
  are  usually  entered as sums of roots of unity, with rational coefficients,
  and  irrational  cyclotomics  are  displayed  in  such  a  way. (For special
  cyclotomics, see [14X18.4[114X.)[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XE(9); E(9)^3; E(6); E(12) / 3;[127X[104X
    [4X[28X-E(9)^4-E(9)^7[128X[104X
    [4X[28XE(3)[128X[104X
    [4X[28X-E(3)^2[128X[104X
    [4X[28X-1/3*E(12)^7[128X[104X
  [4X[32X[104X
  
  [33X[0;0YA  particular basis is used to express cyclotomics, see [14X60.3[114X; note that [10XE(9)[110X
  is [13Xnot[113X a basis element, as the above example shows.[133X
  
  [1X18.1-2 Cyclotomics[101X
  
  [29X[2XCyclotomics[102X[32X global variable
  
  [33X[0;0Yis the domain of all cyclotomics.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XE(9) in Cyclotomics; 37 in Cyclotomics; true in Cyclotomics;[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xfalse[128X[104X
  [4X[32X[104X
  
  [33X[0;0YAs  the cyclotomics are field elements, the usual arithmetic operators [10X+[110X, [10X-[110X,
  [10X*[110X  and [10X/[110X (and [10X^[110X to take powers by integers) are applicable. Note that [10X^[110X does
  [13Xnot[113X  denote  the  conjugation  of  group  elements, so it is [13Xnot[113X possible to
  explicitly  construct  groups  of  cyclotomics.  (However, it is possible to
  compute  the  inverse and the multiplicative order of a nonzero cyclotomic.)
  Also,  taking  the  [22Xk[122X-th  power  of  a  root  of  unity  [22Xz[122X  defines a Galois
  automorphism  if  and  only  if [22Xk[122X is coprime to the conductor (see [2XConductor[102X
  ([14X18.1-7[114X)) of [22Xz[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XE(5) + E(3); (E(5) + E(5)^4) ^ 2; E(5) / E(3); E(5) * E(3);[127X[104X
    [4X[28X-E(15)^2-2*E(15)^8-E(15)^11-E(15)^13-E(15)^14[128X[104X
    [4X[28X-2*E(5)-E(5)^2-E(5)^3-2*E(5)^4[128X[104X
    [4X[28XE(15)^13[128X[104X
    [4X[28XE(15)^8[128X[104X
    [4X[25Xgap>[125X [27XOrder( E(5) ); Order( 1+E(5) );[127X[104X
    [4X[28X5[128X[104X
    [4X[28Xinfinity[128X[104X
  [4X[32X[104X
  
  [1X18.1-3 IsCyclotomic[101X
  
  [29X[2XIsCyclotomic[102X( [3Xobj[103X ) [32X Category
  [29X[2XIsCyc[102X( [3Xobj[103X ) [32X Category
  
  [33X[0;0YEvery   object   in  the  family  [10XCyclotomicsFamily[110X  lies  in  the  category
  [2XIsCyclotomic[102X.  This  covers  integers,  rationals,  proper  cyclotomics, the
  object  [2Xinfinity[102X  ([14X18.2-1[114X), and unknowns (see Chapter [14X74[114X). All these objects
  except  [2Xinfinity[102X  ([14X18.2-1[114X)  and  unknowns  lie  also  in the category [2XIsCyc[102X,
  [2Xinfinity[102X ([14X18.2-1[114X) lies in (and can be detected from) the category [2XIsInfinity[102X
  ([14X18.2-1[114X), and unknowns lie in [2XIsUnknown[102X ([14X74.1-3[114X).[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XIsCyclotomic(0); IsCyclotomic(1/2*E(3)); IsCyclotomic( infinity );[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[25Xgap>[125X [27XIsCyc(0); IsCyc(1/2*E(3)); IsCyc( infinity );[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xfalse[128X[104X
  [4X[32X[104X
  
  [1X18.1-4 IsIntegralCyclotomic[101X
  
  [29X[2XIsIntegralCyclotomic[102X( [3Xobj[103X ) [32X property
  
  [33X[0;0YA  cyclotomic is called [13Xintegral[113X or a [13Xcyclotomic integer[113X if all coefficients
  of  its  minimal  polynomial  over  the  rationals  are  integers. Since the
  underlying  basis  of  the  external  representation  of  cyclotomics  is an
  integral   basis  (see [14X60.3[114X),  the  subring  of  cyclotomic  integers  in  a
  cyclotomic  field  is  formed  by  those  cyclotomics for which the external
  representation  is a list of integers. For example, square roots of integers
  are  cyclotomic  integers  (see [14X18.4[114X),  any  root  of  unity is a cyclotomic
  integer,  character values are always cyclotomic integers, but all rationals
  which are not integers are not cyclotomic integers.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xr:= ER( 5 );               # The square root of 5 ...[127X[104X
    [4X[28XE(5)-E(5)^2-E(5)^3+E(5)^4[128X[104X
    [4X[25Xgap>[125X [27XIsIntegralCyclotomic( r ); # ... is a cyclotomic integer.[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[25Xgap>[125X [27Xr2:= 1/2 * r;              # This is not a cyclotomic integer, ...[127X[104X
    [4X[28X1/2*E(5)-1/2*E(5)^2-1/2*E(5)^3+1/2*E(5)^4[128X[104X
    [4X[25Xgap>[125X [27XIsIntegralCyclotomic( r2 );[127X[104X
    [4X[28Xfalse[128X[104X
    [4X[25Xgap>[125X [27Xr3:= 1/2 * r - 1/2;        # ... but this is one.[127X[104X
    [4X[28XE(5)+E(5)^4[128X[104X
    [4X[25Xgap>[125X [27XIsIntegralCyclotomic( r3 );[127X[104X
    [4X[28Xtrue[128X[104X
  [4X[32X[104X
  
  [1X18.1-5 Int[101X
  
  [29X[2XInt[102X( [3Xcyc[103X ) [32X method
  
  [33X[0;0YThe  operation  [2XInt[102X  can  be  used  to  find a cyclotomic integer near to an
  arbitrary cyclotomic, by applying [2XInt[102X ([14X14.2-3[114X) to the coefficients.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XInt( E(5)+1/2*E(5)^2 ); Int( 2/3*E(7)-3/2*E(4) );[127X[104X
    [4X[28XE(5)[128X[104X
    [4X[28X-E(4)[128X[104X
  [4X[32X[104X
  
  [1X18.1-6 String[101X
  
  [29X[2XString[102X( [3Xcyc[103X ) [32X method
  
  [33X[0;0YThe  operation [2XString[102X returns for a cyclotomic [3Xcyc[103X a string corresponding to
  the way the cyclotomic is printed by [2XViewObj[102X ([14X6.3-5[114X) and [2XPrintObj[102X ([14X6.3-5[114X).[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XString( E(5)+1/2*E(5)^2 ); String( 17/3 );[127X[104X
    [4X[28X"E(5)+1/2*E(5)^2"[128X[104X
    [4X[28X"17/3"[128X[104X
  [4X[32X[104X
  
  [1X18.1-7 Conductor[101X
  
  [29X[2XConductor[102X( [3Xcyc[103X ) [32X attribute
  [29X[2XConductor[102X( [3XC[103X ) [32X attribute
  
  [33X[0;0YFor  an  element  [3Xcyc[103X  of a cyclotomic field, [2XConductor[102X returns the smallest
  integer  [22Xn[122X  such  that  [3Xcyc[103X is contained in the [22Xn[122X-th cyclotomic field. For a
  collection  [3XC[103X  of  cyclotomics (for example a dense list of cyclotomics or a
  field  of  cyclotomics),  [2XConductor[102X returns the smallest integer [22Xn[122X such that
  all elements of [3XC[103X are contained in the [22Xn[122X-th cyclotomic field.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XConductor( 0 ); Conductor( E(10) ); Conductor( E(12) );[127X[104X
    [4X[28X1[128X[104X
    [4X[28X5[128X[104X
    [4X[28X12[128X[104X
  [4X[32X[104X
  
  [1X18.1-8 AbsoluteValue[101X
  
  [29X[2XAbsoluteValue[102X( [3Xcyc[103X ) [32X attribute
  
  [33X[0;0Yreturns  the  absolute  value of a cyclotomic number [3Xcyc[103X. At the moment only
  methods for rational numbers exist.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XAbsoluteValue(-3);[127X[104X
    [4X[28X3[128X[104X
  [4X[32X[104X
  
  [1X18.1-9 RoundCyc[101X
  
  [29X[2XRoundCyc[102X( [3Xcyc[103X ) [32X operation
  
  [33X[0;0Yis  a  cyclotomic  integer [22Xz[122X (see [2XIsIntegralCyclotomic[102X ([14X18.1-4[114X)) near to the
  cyclotomic  [3Xcyc[103X in the following sense: Let [10Xc[110X be the [22Xi[122X-th coefficient in the
  external  representation  (see [2XCoeffsCyc[102X  ([14X18.1-10[114X))  of  [3Xcyc[103X. Then the [22Xi[122X-th
  coefficient  in the external representation of [22Xz[122X is [10XInt( c + 1/2 )[110X or [10XInt( c
  - 1/2 )[110X, depending on whether [10Xc[110X is nonnegative or negative, respectively.[133X
  
  [33X[0;0YExpressed   in  terms  of  the  Zumbroich  basis  (see [14X60.3[114X),  rounding  the
  coefficients  of  [3Xcyc[103X  w.r.t. this  basis  to the nearest integer yields the
  coefficients of [22Xz[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XRoundCyc( E(5)+1/2*E(5)^2 ); RoundCyc( 2/3*E(7)+3/2*E(4) );[127X[104X
    [4X[28XE(5)+E(5)^2[128X[104X
    [4X[28X-2*E(28)^3+E(28)^4-2*E(28)^11-2*E(28)^15-2*E(28)^19-2*E(28)^23[128X[104X
    [4X[28X -2*E(28)^27[128X[104X
  [4X[32X[104X
  
  [1X18.1-10 CoeffsCyc[101X
  
  [29X[2XCoeffsCyc[102X( [3Xcyc[103X, [3XN[103X ) [32X function
  
  [33X[0;0YLet  [3Xcyc[103X  be a cyclotomic with conductor [22Xn[122X (see [2XConductor[102X ([14X18.1-7[114X)). If [3XN[103X is
  not  a  multiple  of  [22Xn[122X  then  [2XCoeffsCyc[102X  returns [9Xfail[109X because [3Xcyc[103X cannot be
  expressed  in  terms  of  [3XN[103X-th roots of unity. Otherwise [2XCoeffsCyc[102X returns a
  list  of length [3XN[103X with entry at position [22Xj[122X equal to the coefficient of [22Xexp(2
  π  i  (j-1)/[3XN[103X)[122X  if this root belongs to the [3XN[103X-th Zumbroich basis (see [14X60.3[114X),
  and  equal  to  zero otherwise. So we have [3Xcyc[103X = [10XCoeffsCyc([110X [3Xcyc[103X, [3XN[103X [10X) * List(
  [1..[110X[3XN[103X[10X], j -> E([110X[3XN[103X[10X)^(j-1) )[110X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xcyc:= E(5)+E(5)^2;[127X[104X
    [4X[28XE(5)+E(5)^2[128X[104X
    [4X[25Xgap>[125X [27XCoeffsCyc( cyc, 5 );  CoeffsCyc( cyc, 15 );  CoeffsCyc( cyc, 7 );[127X[104X
    [4X[28X[ 0, 1, 1, 0, 0 ][128X[104X
    [4X[28X[ 0, -1, 0, 0, 0, 0, 0, 0, -1, 0, 0, -1, 0, -1, 0 ][128X[104X
    [4X[28Xfail[128X[104X
  [4X[32X[104X
  
  [1X18.1-11 DenominatorCyc[101X
  
  [29X[2XDenominatorCyc[102X( [3Xcyc[103X ) [32X function
  
  [33X[0;0YFor  a  cyclotomic  number  [3Xcyc[103X  (see [2XIsCyclotomic[102X  ([14X18.1-3[114X)), this function
  returns  the  smallest  positive integer [22Xn[122X such that [22Xn[122X[10X * [110X[3Xcyc[103X is a cyclotomic
  integer  (see [2XIsIntegralCyclotomic[102X  ([14X18.1-4[114X)). For rational numbers [3Xcyc[103X, the
  result is the same as that of [2XDenominatorRat[102X ([14X17.2-5[114X).[133X
  
  [1X18.1-12 ExtRepOfObj[101X
  
  [29X[2XExtRepOfObj[102X( [3Xcyc[103X ) [32X method
  
  [33X[0;0YThe  external  representation  of  a  cyclotomic  [3Xcyc[103X  with conductor [22Xn[122X (see
  [2XConductor[102X  ([14X18.1-7[114X) is the list returned by [2XCoeffsCyc[102X ([14X18.1-10[114X), called with
  [3Xcyc[103X and [22Xn[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XExtRepOfObj( E(5) ); CoeffsCyc( E(5), 5 );[127X[104X
    [4X[28X[ 0, 1, 0, 0, 0 ][128X[104X
    [4X[28X[ 0, 1, 0, 0, 0 ][128X[104X
    [4X[25Xgap>[125X [27XCoeffsCyc( E(5), 15 );[127X[104X
    [4X[28X[ 0, 0, 0, 0, 0, 0, 0, 0, -1, 0, 0, 0, 0, -1, 0 ][128X[104X
  [4X[32X[104X
  
  [1X18.1-13 DescriptionOfRootOfUnity[101X
  
  [29X[2XDescriptionOfRootOfUnity[102X( [3Xroot[103X ) [32X function
  
  [33X[0;0YGiven  a  cyclotomic  [3Xroot[103X  that is known to be a root of unity (this is [13Xnot[113X
  checked),  [2XDescriptionOfRootOfUnity[102X  returns  a  list  [22X[  n,  e ][122X of coprime
  positive integers such that [3Xroot[103X [22X=[122X [10XE[110X[22X(n)^e[122X holds.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XE(9);  DescriptionOfRootOfUnity( E(9) );[127X[104X
    [4X[28X-E(9)^4-E(9)^7[128X[104X
    [4X[28X[ 9, 1 ][128X[104X
    [4X[25Xgap>[125X [27XDescriptionOfRootOfUnity( -E(3) );[127X[104X
    [4X[28X[ 6, 5 ][128X[104X
  [4X[32X[104X
  
  [1X18.1-14 IsGaussInt[101X
  
  [29X[2XIsGaussInt[102X( [3Xx[103X ) [32X function
  
  [33X[0;0Y[2XIsGaussInt[102X   returns   [9Xtrue[109X   if   the   object  [3Xx[103X  is  a  Gaussian  integer
  (see [2XGaussianIntegers[102X  ([14X60.5-1[114X)), and [9Xfalse[109X otherwise. Gaussian integers are
  of the form [22Xa + b[122X[10X*E(4)[110X, where [22Xa[122X and [22Xb[122X are integers.[133X
  
  [1X18.1-15 IsGaussRat[101X
  
  [29X[2XIsGaussRat[102X( [3Xx[103X ) [32X function
  
  [33X[0;0Y[2XIsGaussRat[102X   returns   [9Xtrue[109X   if   the  object  [3Xx[103X  is  a  Gaussian  rational
  (see [2XGaussianRationals[102X  ([14X60.1-3[114X)),  and  [9Xfalse[109X otherwise. Gaussian rationals
  are of the form [22Xa + b[122X[10X*E(4)[110X, where [22Xa[122X and [22Xb[122X are rationals.[133X
  
  [1X18.1-16 DefaultField[101X
  
  [29X[2XDefaultField[102X( [3Xlist[103X ) [32X function
  
  [33X[0;0Y[2XDefaultField[102X  for  cyclotomics  is defined to return the smallest [13Xcyclotomic[113X
  field containing the given elements.[133X
  
  [33X[0;0YNote  that  [2XField[102X  ([14X58.1-3[114X)  returns the smallest field containing all given
  elements,  which  need  not be a cyclotomic field. In both cases, the fields
  represent vector spaces over the rationals (see [14X60.3[114X).[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XField( E(5)+E(5)^4 );  DefaultField( E(5)+E(5)^4 );[127X[104X
    [4X[28XNF(5,[ 1, 4 ])[128X[104X
    [4X[28XCF(5)[128X[104X
  [4X[32X[104X
  
  
  [1X18.2 [33X[0;0YInfinity and negative Infinity[133X[101X
  
  [1X18.2-1 IsInfinity[101X
  
  [29X[2XIsInfinity[102X( [3Xobj[103X ) [32X Category
  [29X[2XIsNegInfinity[102X( [3Xobj[103X ) [32X Category
  [29X[2Xinfinity[102X[32X global variable
  [29X[2X-infinity[102X[32X global variable
  
  [33X[0;0Y[2Xinfinity[102X   and   [2X-infinity[102X   are   special   [5XGAP[105X   objects   that   lie   in
  [10XCyclotomicsFamily[110X. They are larger or smaller than all other objects in this
  family  respectively.  [2Xinfinity[102X is mainly used as return value of operations
  such  as  [2XSize[102X  ([14X30.4-6[114X)  and  [2XDimension[102X  ([14X57.3-3[114X) for infinite and infinite
  dimensional domains, respectively.[133X
  
  [33X[0;0YSome  arithmetic operations are provided for convenience when using [2Xinfinity[102X
  and [2X-infinity[102X as top and bottom element respectively.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27X-infinity + 1;[127X[104X
    [4X[28X-infinity[128X[104X
    [4X[25Xgap>[125X [27Xinfinity + infinity;[127X[104X
    [4X[28Xinfinity[128X[104X
  [4X[32X[104X
  
  [33X[0;0YOften  it  is  useful  to  distinguish [2Xinfinity[102X from [21Xproper[121X cyclotomics. For
  that,  [2Xinfinity[102X  lies  in the category [2XIsInfinity[102X but not in [2XIsCyc[102X ([14X18.1-3[114X),
  and  the  other  cyclotomics  lie  in the category [2XIsCyc[102X ([14X18.1-3[114X) but not in
  [2XIsInfinity[102X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xs:= Size( Rationals );[127X[104X
    [4X[28Xinfinity[128X[104X
    [4X[25Xgap>[125X [27Xs = infinity; IsCyclotomic( s ); IsCyc( s ); IsInfinity( s );[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xfalse[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[25Xgap>[125X [27Xs in Rationals; s > 17;[127X[104X
    [4X[28Xfalse[128X[104X
    [4X[28Xtrue[128X[104X
    [4X[25Xgap>[125X [27XSet( [ s, 2, s, E(17), s, 19 ] );[127X[104X
    [4X[28X[ 2, 19, E(17), infinity ][128X[104X
  [4X[32X[104X
  
  
  [1X18.3 [33X[0;0YComparisons of Cyclotomics[133X[101X
  
  [33X[0;0YTo  compare  cyclotomics, the operators [10X<[110X, [10X<=[110X, [10X=[110X, [10X>=[110X, [10X>[110X, and [10X<>[110X can be used,
  the  result  will be [9Xtrue[109X if the first operand is smaller, smaller or equal,
  equal,  larger  or  equal,  larger,  or  unequal,  respectively,  and  [9Xfalse[109X
  otherwise.[133X
  
  [33X[0;0YCyclotomics  are  ordered  as follows: The relation between rationals is the
  natural one, rationals are smaller than irrational cyclotomics, and [2Xinfinity[102X
  ([14X18.2-1[114X)  is  the  largest  cyclotomic.  For two irrational cyclotomics with
  different   conductors  (see  [2XConductor[102X  ([14X18.1-7[114X)),  the  one  with  smaller
  conductor  is  regarded  as  smaller.  Two  irrational cyclotomics with same
  conductor  are  compared  via their external representation (see [2XExtRepOfObj[102X
  ([14X18.1-12[114X)).[133X
  
  [33X[0;0YFor comparisons of cyclotomics and other [5XGAP[105X objects, see Section [14X4.12[114X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XE(5) < E(6);      # the latter value has conductor 3[127X[104X
    [4X[28Xfalse[128X[104X
    [4X[25Xgap>[125X [27XE(3) < E(3)^2;    # both have conductor 3, compare the ext. repr.[127X[104X
    [4X[28Xfalse[128X[104X
    [4X[25Xgap>[125X [27X3 < E(3); E(5) < E(7);[127X[104X
    [4X[28Xtrue[128X[104X
    [4X[28Xtrue[128X[104X
  [4X[32X[104X
  
  
  [1X18.4 [33X[0;0YATLAS Irrationalities[133X[101X
  
  
  [1X18.4-1 [33X[0;0YEB, EC, [22X...[122X[101X[1X, EH[133X[101X
  
  [29X[2XEB[102X( [3XN[103X ) [32X function
  [29X[2XEC[102X( [3XN[103X ) [32X function
  [29X[2XED[102X( [3XN[103X ) [32X function
  [29X[2XEE[102X( [3XN[103X ) [32X function
  [29X[2XEF[102X( [3XN[103X ) [32X function
  [29X[2XEG[102X( [3XN[103X ) [32X function
  [29X[2XEH[102X( [3XN[103X ) [32X function
  
  [33X[0;0YFor  a  positive  integer  [3XN[103X,  let  [22Xz  =[122X  [10XE([110X[3XN[103X[10X)[110X [22X= exp(2 π i/[3XN[103X)[122X. The following
  so-called  [13Xatomic irrationalities[113X (see [CCNPW85, Chapter 7, Section 10]) can
  be  entered  using  functions.  (Note  that  the  values  are  not necessary
  irrational.)[133X
  
        [10XEB([110X[3XN[103X[10X)[110X   =   [22Xb_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^2} ) / 2[122X   ,   [22X[3XN[103X ≡ 1 mod 2[122X
        [10XEC([110X[3XN[103X[10X)[110X   =   [22Xc_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^3} ) / 3[122X   ,   [22X[3XN[103X ≡ 1 mod 3[122X
        [10XED([110X[3XN[103X[10X)[110X   =   [22Xd_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^4} ) / 4[122X   ,   [22X[3XN[103X ≡ 1 mod 4[122X
        [10XEE([110X[3XN[103X[10X)[110X   =   [22Xe_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^5} ) / 5[122X   ,   [22X[3XN[103X ≡ 1 mod 5[122X
        [10XEF([110X[3XN[103X[10X)[110X   =   [22Xf_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^6} ) / 6[122X   ,   [22X[3XN[103X ≡ 1 mod 6[122X
        [10XEG([110X[3XN[103X[10X)[110X   =   [22Xg_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^7} ) / 7[122X   ,   [22X[3XN[103X ≡ 1 mod 7[122X
        [10XEH([110X[3XN[103X[10X)[110X   =   [22Xh_[3XN[103X[122X   =   [22X( ∑_{j = 1}^{[3XN[103X-1} z^{j^8} ) / 8[122X   ,   [22X[3XN[103X ≡ 1 mod 8[122X
  
  [33X[0;0Y(Note that in [10XEC([110X[3XN[103X[10X)[110X, [22X...[122X, [10XEH([110X[3XN[103X[10X)[110X, [3XN[103X must be a prime.)[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XEB(5);  EB(9);[127X[104X
    [4X[28XE(5)+E(5)^4[128X[104X
    [4X[28X1[128X[104X
  [4X[32X[104X
  
  
  [1X18.4-2 [33X[0;0YEI and ER[133X[101X
  
  [29X[2XEI[102X( [3XN[103X ) [32X function
  [29X[2XER[102X( [3XN[103X ) [32X function
  
  [33X[0;0YFor  a  rational  number  [3XN[103X, [2XER[102X returns the square root [22Xsqrt{[3XN[103X}[122X of [3XN[103X, and [2XEI[102X
  returns  [22Xsqrt{-[3XN[103X}[122X.  By  the  chosen  embedding of cyclotomic fields into the
  complex  numbers,  [2XER[102X returns the positive square root if [3XN[103X is positive, and
  if  [3XN[103X  is  negative  then  [10XER([110X[3XN[103X[10X) = EI(-[110X[3XN[103X[10X)[110X holds. In any case, [10XEI([110X[3XN[103X[10X) = E(4) *
  ER([110X[3XN[103X[10X)[110X.[133X
  
  [33X[0;0Y[2XER[102X  is  installed  as  method for the operation [2XSqrt[102X ([14X31.12-5[114X), for rational
  argument.[133X
  
  [33X[0;0YFrom a theorem of Gauss we know that [22Xb_[3XN[103X =[122X[133X
  
        [22X(-1 + sqrt{[3XN[103X}) / 2[122X     if   [22X[3XN[103X ≡ 1 mod 4[122X 
        [22X(-1 + i sqrt{[3XN[103X}) / 2[122X   if   [22X[3XN[103X ≡ -1 mod 4[122X
  
  [33X[0;0YSo [22Xsqrt{[3XN[103X}[122X can be computed from [22Xb_[3XN[103X[122X, see [2XEB[102X ([14X18.4-1[114X).[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XER(3); EI(3);[127X[104X
    [4X[28X-E(12)^7+E(12)^11[128X[104X
    [4X[28XE(3)-E(3)^2[128X[104X
  [4X[32X[104X
  
  
  [1X18.4-3 [33X[0;0YEY, EX, [22X...[122X[101X[1X, ES[133X[101X
  
  [29X[2XEY[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEX[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEW[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEV[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEU[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XET[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XES[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  
  [33X[0;0YFor  the  given  integer  [3XN[103X  [22X>  2[122X,  let  [22X[3XN[103X_k[122X  denote  the first integer with
  multiplicative order exactly [22Xk[122X modulo [3XN[103X, chosen in the order of preference[133X
  
  
        [33X[1;6Y[24X[33X[0;0Y1, -1, 2, -2, 3, -3, 4, -4, ... .[133X[124X[133X
  
  
  [33X[0;0YWe define (with [22Xz = exp(2 π i/[3XN[103X)[122X)[133X
  
        [10XEY([110X[3XN[103X[10X)[110X   =   [22Xy_[3XN[103X[122X   =   [22Xz + z^n[122X                                 [22X(n = [3XN[103X_2)[122X
        [10XEX([110X[3XN[103X[10X)[110X   =   [22Xx_[3XN[103X[122X   =   [22Xz + z^n + z^{n^2}[122X                       [22X(n = [3XN[103X_3)[122X
        [10XEW[110X([3XN[103X[10X)[110X   =   [22Xw_[3XN[103X[122X   =   [22Xz + z^n + z^{n^2} + z^{n^3}[122X             [22X(n = [3XN[103X_4)[122X
        [10XEV([110X[3XN[103X[10X)[110X   =   [22Xv_[3XN[103X[122X   =   [22Xz + z^n + z^{n^2} + z^{n^3} + z^{n^4}[122X   [22X(n = [3XN[103X_5)[122X
        [10XEU([110X[3XN[103X[10X)[110X   =   [22Xu_[3XN[103X[122X   =   [22Xz + z^n + z^{n^2} + ... + z^{n^5}[122X       [22X(n = [3XN[103X_6)[122X
        [10XET([110X[3XN[103X[10X)[110X   =   [22Xt_[3XN[103X[122X   =   [22Xz + z^n + z^{n^2} + ... + z^{n^6}[122X       [22X(n = [3XN[103X_7)[122X
        [10XES([110X[3XN[103X[10X)[110X   =   [22Xs_[3XN[103X[122X   =   [22Xz + z^n + z^{n^2} + ... + z^{n^7}[122X       [22X(n = [3XN[103X_8)[122X
  
  [33X[0;0YFor the two-argument versions of the functions, see Section [2XNK[102X ([14X18.4-5[114X).[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XEY(5);[127X[104X
    [4X[28XE(5)+E(5)^4[128X[104X
    [4X[25Xgap>[125X [27XEW(16,3); EW(17,2);[127X[104X
    [4X[28X0[128X[104X
    [4X[28XE(17)+E(17)^4+E(17)^13+E(17)^16[128X[104X
  [4X[32X[104X
  
  
  [1X18.4-4 [33X[0;0YEM, EL, [22X...[122X[101X[1X, EJ[133X[101X
  
  [29X[2XEM[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEL[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEK[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  [29X[2XEJ[102X( [3XN[103X[, [3Xd[103X] ) [32X function
  
  [33X[0;0YLet [3XN[103X be an integer, [3XN[103X [22X> 2[122X. We define (with [22Xz = exp(2 π i/[3XN[103X)[122X)[133X
  
        [10XEM([110X[3XN[103X[10X)[110X   =   [22Xm_[3XN[103X[122X   =   [22Xz - z^n[122X                       [22X(n = [3XN[103X_2)[122X
        [10XEL([110X[3XN[103X[10X)[110X   =   [22Xl_[3XN[103X[122X   =   [22Xz - z^n + z^{n^2} - z^{n^3}[122X   [22X(n = [3XN[103X_4)[122X
        [10XEK([110X[3XN[103X[10X)[110X   =   [22Xk_[3XN[103X[122X   =   [22Xz - z^n + ... - z^{n^5}[122X       [22X(n = [3XN[103X_6)[122X
        [10XEJ([110X[3XN[103X[10X)[110X   =   [22Xj_[3XN[103X[122X   =   [22Xz - z^n + ... - z^{n^7}[122X       [22X(n = [3XN[103X_8)[122X
  
  [33X[0;0YFor the two-argument versions of the functions, see Section [2XNK[102X ([14X18.4-5[114X).[133X
  
  [1X18.4-5 NK[101X
  
  [29X[2XNK[102X( [3XN[103X, [3Xk[103X, [3Xd[103X ) [32X function
  
  [33X[0;0YLet  [22X[3XN[103X_[3Xk[103X^([3Xd[103X)[122X  be  the  [22X([3Xd[103X+1)[122X-th  integer with multiplicative order exactly [3Xk[103X
  modulo  [3XN[103X,  chosen  in the order of preference defined in Section [14X18.4-3[114X; [2XNK[102X
  returns  [22X[3XN[103X_[3Xk[103X^([3Xd[103X)[122X;  if  there  is no integer with the required multiplicative
  order, [2XNK[102X returns [9Xfail[109X.[133X
  
  [33X[0;0YWe write [22X[3XN[103X_[3Xk[103X = [3XN[103X_[3Xk[103X^(0), [3XN[103X_[3Xk[103X^' = [3XN[103X_[3Xk[103X^(1), [3XN[103X_[3Xk[103X^'' = [3XN[103X_[3Xk[103X^(2)[122X and so on.[133X
  
  [33X[0;0YThe algebraic numbers[133X
  
  
        [33X[1;6Y[24X[33X[0;0Yy_[3XN[103X^' = y_[3XN[103X^(1), y_[3XN[103X^'' = y_[3XN[103X^(2), ..., x_[3XN[103X^', x_[3XN[103X^'', ..., j_[3XN[103X^',
        j_[3XN[103X^'', ...[133X[124X[133X
  
  
  [33X[0;0Yare  obtained on replacing [22X[3XN[103X_[3Xk[103X[122X in the definitions in the sections [14X18.4-3[114X and
  [14X18.4-4[114X by [22X[3XN[103X_[3Xk[103X^', [3XN[103X_[3Xk[103X^'', ...[122X; they can be entered as[133X
  
        [10XEY([110X[3XN[103X,[3Xd[103X[10X)[110X    =    [22Xy_[3XN[103X^([3Xd[103X)[122X
        [10XEX([110X[3XN[103X,[3Xd[103X[10X)[110X    =    [22Xx_[3XN[103X^([3Xd[103X)[122X
                  [22X...[122X          
        [10XEJ([110X[3XN[103X,[3Xd[103X[10X)[110X    =    [22Xj_[3XN[103X^([3Xd[103X)[122X
  
  [1X18.4-6 AtlasIrrationality[101X
  
  [29X[2XAtlasIrrationality[102X( [3Xirratname[103X ) [32X function
  
  [33X[0;0YLet  [3Xirratname[103X  be  a  string that describes an irrational value as a linear
  combination  in  terms  of  the  atomic  irrationalities  introduced  in the
  sections [14X18.4-1[114X, [14X18.4-2[114X, [14X18.4-3[114X, [14X18.4-4[114X. These irrational values are defined
  in [CCNPW85, Chapter 6, Section 10], and the following description is mainly
  copied  from  there.  If  [22Xq_N[122X  is  such  a  value (e.g. [22Xy_24^''[122X) then linear
  combinations  of  algebraic  conjugates  of  [22Xq_N[122X  are  abbreviated as in the
  following examples:[133X
  
        [10X2qN+3&5-4&7+&9[110X   means   [22X2 q_N + 3 q_N^{*5} - 4 q_N^{*7} + q_N^{*9}[122X            
        [10X4qN&3&5&7-3&4[110X    means   [22X4 (q_N + q_N^{*3} + q_N^{*5} + q_N^{*7}) - 3 q_N^{*11}[122X
        [10X4qN*3&5+&7[110X       means   [22X4 (q_N^{*3} + q_N^{*5}) + q_N^{*7}[122X                    
  
  [33X[0;0YTo  explain the [21Xampersand[121X syntax in general we remark that [21X&k[121X is interpreted
  as  [22Xq_N^{*k}[122X, where [22Xq_N[122X is the most recently named atomic irrationality, and
  that the scope of any premultiplying coefficient is broken by a [22X+[122X or [22X-[122X sign,
  but  not  by [22X&[122X or [22X*k[122X. The algebraic conjugations indicated by the ampersands
  apply  directly  to  the [13Xatomic[113X irrationality [22Xq_N[122X, even when, as in the last
  example, [22Xq_N[122X first appears with another conjugacy [22X*k[122X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "b7*3" );[127X[104X
    [4X[28XE(7)^3+E(7)^5+E(7)^6[128X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "y'''24" );[127X[104X
    [4X[28XE(24)-E(24)^19[128X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "-3y'''24*13&5" );[127X[104X
    [4X[28X3*E(8)-3*E(8)^3[128X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "3y'''24*13-2&5" );[127X[104X
    [4X[28X-3*E(24)-2*E(24)^11+2*E(24)^17+3*E(24)^19[128X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "3y'''24*13-&5" );[127X[104X
    [4X[28X-3*E(24)-E(24)^11+E(24)^17+3*E(24)^19[128X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "3y'''24*13-4&5&7" );[127X[104X
    [4X[28X-7*E(24)-4*E(24)^11+4*E(24)^17+7*E(24)^19[128X[104X
    [4X[25Xgap>[125X [27XAtlasIrrationality( "3y'''24&7" );[127X[104X
    [4X[28X6*E(24)-6*E(24)^19[128X[104X
  [4X[32X[104X
  
  
  [1X18.5 [33X[0;0YGalois Conjugacy of Cyclotomics[133X[101X
  
  [1X18.5-1 GaloisCyc[101X
  
  [29X[2XGaloisCyc[102X( [3Xcyc[103X, [3Xk[103X ) [32X operation
  [29X[2XGaloisCyc[102X( [3Xlist[103X, [3Xk[103X ) [32X operation
  
  [33X[0;0YFor  a  cyclotomic  [3Xcyc[103X  and  an integer [3Xk[103X, [2XGaloisCyc[102X returns the cyclotomic
  obtained by raising the roots of unity in the Zumbroich basis representation
  of  [3Xcyc[103X to the [3Xk[103X-th power. If [3Xk[103X is coprime to the integer [22Xn[122X, [10XGaloisCyc( ., [3Xk[103X[10X
  )[110X  acts as a Galois automorphism of the [22Xn[122X-th cyclotomic field (see [14X60.4[114X); to
  get the Galois automorphisms themselves, use [2XGaloisGroup[102X ([14X58.3-1[114X).[133X
  
  [33X[0;0YThe  [13Xcomplex  conjugate[113X  of  [3Xcyc[103X  is [10XGaloisCyc( [3Xcyc[103X[10X, -1 )[110X, which can also be
  computed using [2XComplexConjugate[102X ([14X18.5-2[114X).[133X
  
  [33X[0;0YFor  a  list  or  matrix  [3Xlist[103X  of  cyclotomics,  [2XGaloisCyc[102X returns the list
  obtained by applying [2XGaloisCyc[102X to the entries of [3Xlist[103X.[133X
  
  [1X18.5-2 ComplexConjugate[101X
  
  [29X[2XComplexConjugate[102X( [3Xz[103X ) [32X attribute
  [29X[2XRealPart[102X( [3Xz[103X ) [32X attribute
  [29X[2XImaginaryPart[102X( [3Xz[103X ) [32X attribute
  
  [33X[0;0YFor  a  cyclotomic  number  [3Xz[103X,  [2XComplexConjugate[102X returns [10XGaloisCyc( [3Xz[103X[10X, -1 )[110X,
  see [2XGaloisCyc[102X  ([14X18.5-1[114X). For a quaternion [22X[3Xz[103X = c_1 e + c_2 i + c_3 j + c_4 k[122X,
  [2XComplexConjugate[102X  returns  [22Xc_1  e  - c_2 i - c_3 j - c_4 k[122X, see [2XIsQuaternion[102X
  ([14X62.8-8[114X).[133X
  
  [33X[0;0YWhen  [2XComplexConjugate[102X  is called with a list then the result is the list of
  return  values of [2XComplexConjugate[102X for the list entries in the corresponding
  positions.[133X
  
  [33X[0;0YWhen  [2XComplexConjugate[102X  is  defined  for  an  object  [3Xz[103X  then  [2XRealPart[102X  and
  [2XImaginaryPart[102X   return   [10X([3Xz[103X[10X   +   ComplexConjugate(  [3Xz[103X[10X  ))  /  2[110X  and  [10X([3Xz[103X[10X  -
  ComplexConjugate(   [3Xz[103X[10X   ))   /  2  i[110X,  respectively,  where  [10Xi[110X  denotes  the
  corresponding imaginary unit.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XGaloisCyc( E(5) + E(5)^4, 2 );[127X[104X
    [4X[28XE(5)^2+E(5)^3[128X[104X
    [4X[25Xgap>[125X [27XGaloisCyc( E(5), -1 );           # the complex conjugate[127X[104X
    [4X[28XE(5)^4[128X[104X
    [4X[25Xgap>[125X [27XGaloisCyc( E(5) + E(5)^4, -1 );  # this value is real[127X[104X
    [4X[28XE(5)+E(5)^4[128X[104X
    [4X[25Xgap>[125X [27XGaloisCyc( E(15) + E(15)^4, 3 );[127X[104X
    [4X[28XE(5)+E(5)^4[128X[104X
    [4X[25Xgap>[125X [27XComplexConjugate( E(7) );[127X[104X
    [4X[28XE(7)^6[128X[104X
  [4X[32X[104X
  
  [1X18.5-3 StarCyc[101X
  
  [29X[2XStarCyc[102X( [3Xcyc[103X ) [32X function
  
  [33X[0;0YIf  the  cyclotomic [3Xcyc[103X is an irrational element of a quadratic extension of
  the  rationals  then [2XStarCyc[102X returns the unique Galois conjugate of [3Xcyc[103X that
  is  different  from  [3Xcyc[103X, otherwise [9Xfail[109X is returned. In the first case, the
  return value is often called [3Xcyc[103X[22X*[122X (see [14X71.13[114X).[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XStarCyc( EB(5) ); StarCyc( E(5) );[127X[104X
    [4X[28XE(5)^2+E(5)^3[128X[104X
    [4X[28Xfail[128X[104X
  [4X[32X[104X
  
  [1X18.5-4 Quadratic[101X
  
  [29X[2XQuadratic[102X( [3Xcyc[103X ) [32X function
  
  [33X[0;0YLet  [3Xcyc[103X be a cyclotomic integer that lies in a quadratic extension field of
  the  rationals. Then we have [3Xcyc[103X[22X= (a + b sqrt{n}) / d[122X, for integers [22Xa[122X, [22Xb[122X, [22Xn[122X,
  [22Xd[122X,  such  that  [22Xd[122X is either [22X1[122X or [22X2[122X. In this case, [2XQuadratic[102X returns a record
  with  the  components  [10Xa[110X,  [10Xb[110X, [10Xroot[110X, [10Xd[110X, [10XATLAS[110X, and [10Xdisplay[110X; the values of the
  first  four  are  [22Xa[122X,  [22Xb[122X,  [22Xn[122X,  and  [22Xd[122X,  the [10XATLAS[110X value is a (not necessarily
  shortest)  representation  of  [3Xcyc[103X  in  terms  of  the [5XAtlas[105X irrationalities
  [22Xb_{|n|}[122X,  [22Xi_{|n|}[122X, [22Xr_{|n|}[122X, and the [10Xdisplay[110X value is a string that expresses
  [3Xcyc[103X in [5XGAP[105X notation, corresponding to the value of the [10XATLAS[110X component.[133X
  
  [33X[0;0YIf  [3Xcyc[103X is not a cyclotomic integer or does not lie in a quadratic extension
  field of the rationals then [9Xfail[109X is returned.[133X
  
  [33X[0;0YIf the denominator [22Xd[122X is [22X2[122X then necessarily [22Xn[122X is congruent to [22X1[122X modulo [22X4[122X, and
  [22Xr_n[122X,  [22Xi_n[122X are not possible; we have [10X[3Xcyc[103X[10X = x + y * EB( root )[110X with [10Xy = b[110X, [10Xx =
  ( a + b ) / 2[110X.[133X
  
  [33X[0;0YIf  [22Xd  =  1[122X, we have the possibilities [22Xi_{|n|}[122X for [22Xn < -1[122X, [22Xa + b * i[122X for [22Xn =
  -1[122X, [22Xa + b * r_n[122X for [22Xn > 0[122X. Furthermore if [22Xn[122X is congruent to [22X1[122X modulo [22X4[122X, also
  [3Xcyc[103X  [22X=  (a+b) + 2 * b * b_{|n|}[122X is possible; the shortest string of these is
  taken as the value for the component [10XATLAS[110X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XQuadratic( EB(5) ); Quadratic( EB(27) );[127X[104X
    [4X[28Xrec( ATLAS := "b5", a := -1, b := 1, d := 2, [128X[104X
    [4X[28X  display := "(-1+Sqrt(5))/2", root := 5 )[128X[104X
    [4X[28Xrec( ATLAS := "1+3b3", a := -1, b := 3, d := 2, [128X[104X
    [4X[28X  display := "(-1+3*Sqrt(-3))/2", root := -3 )[128X[104X
    [4X[25Xgap>[125X [27XQuadratic(0); Quadratic( E(5) );[127X[104X
    [4X[28Xrec( ATLAS := "0", a := 0, b := 0, d := 1, display := "0", root := 1 )[128X[104X
    [4X[28Xfail[128X[104X
  [4X[32X[104X
  
  [1X18.5-5 GaloisMat[101X
  
  [29X[2XGaloisMat[102X( [3Xmat[103X ) [32X attribute
  
  [33X[0;0YLet [3Xmat[103X be a matrix of cyclotomics. [2XGaloisMat[102X calculates the complete orbits
  under  the operation of the Galois group of the (irrational) entries of [3Xmat[103X,
  and  the  permutations of rows corresponding to the generators of the Galois
  group.[133X
  
  [33X[0;0YIf  some rows of [3Xmat[103X are identical, only the first one is considered for the
  permutations, and a warning will be printed.[133X
  
  [33X[0;0Y[2XGaloisMat[102X  returns  a  record  with  the  components  [10Xmat[110X,  [10Xgaloisfams[110X,  and
  [10Xgenerators[110X.[133X
  
  [8X[10Xmat[110X[108X
        [33X[0;6Ya  list with initial segment being the rows of [3Xmat[103X ([13Xnot[113X shallow copies
        of  these  rows); the list consists of full orbits under the action of
        the Galois group of the entries of [3Xmat[103X defined above. The last rows in
        the  list are those not contained in [3Xmat[103X but must be added in order to
        complete  the  orbits; so if the orbits were already complete, [3Xmat[103X and
        [10Xmat[110X have identical rows.[133X
  
  [8X[10Xgaloisfams[110X[108X
        [33X[0;6Ya  list that has the same length as the [10Xmat[110X component, its entries are
        either 1, 0, -1, or lists.[133X
  
        [8X[10Xgaloisfams[i] = 1[110X[108X
              [33X[0;12Ymeans  that [10Xmat[i][110X consists of rationals, i.e., [10X[ mat[i] ][110X forms
              an orbit;[133X
  
        [8X[10Xgaloisfams[i] = -1[110X[108X
              [33X[0;12Ymeans  that  [10Xmat[i][110X  contains unknowns (see Chapter [14X74[114X); in this
              case  [10X[  mat[i]  ][110X  is regarded as an orbit, too, even if [10Xmat[i][110X
              contains irrational entries;[133X
  
        [8X[10Xgaloisfams[i] = [110X[22X[ l_1, l_2 ][122X[108X
              [33X[0;12Y(a  list) means that [10Xmat[i][110X is the first element of its orbit in
              [10Xmat[110X,  [22Xl_1[122X  is the list of positions of rows that form the orbit,
              and  [22Xl_2[122X  is  the list of corresponding Galois automorphisms (as
              exponents,  not  as  functions);  so  we have [10Xmat[110X[22X[ l_1[j] ][k] =[122X
              [10XGaloisCyc( mat[110X[22X[i][k], l_2[j][122X[10X )[110X;[133X
  
        [8X[10Xgaloisfams[i] = 0[110X[108X
              [33X[0;12Ymeans  that  [10Xmat[i][110X  is an element of a nontrivial orbit but not
              the first element of it.[133X
  
  [8X[10Xgenerators[110X[108X
        [33X[0;6Ya  list of permutations generating the permutation group corresponding
        to the action of the Galois group on the rows of [10Xmat[110X.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XGaloisMat( [ [ E(3), E(4) ] ] );[127X[104X
    [4X[28Xrec( galoisfams := [ [ [ 1, 2, 3, 4 ], [ 1, 7, 5, 11 ] ], 0, 0, 0 ], [128X[104X
    [4X[28X  generators := [ (1,2)(3,4), (1,3)(2,4) ], [128X[104X
    [4X[28X  mat := [ [ E(3), E(4) ], [ E(3), -E(4) ], [ E(3)^2, E(4) ], [128X[104X
    [4X[28X      [ E(3)^2, -E(4) ] ] )[128X[104X
    [4X[25Xgap>[125X [27XGaloisMat( [ [ 1, 1, 1 ], [ 1, E(3), E(3)^2 ] ] );[127X[104X
    [4X[28Xrec( galoisfams := [ 1, [ [ 2, 3 ], [ 1, 2 ] ], 0 ], [128X[104X
    [4X[28X  generators := [ (2,3) ], [128X[104X
    [4X[28X  mat := [ [ 1, 1, 1 ], [ 1, E(3), E(3)^2 ], [ 1, E(3)^2, E(3) ] ] )[128X[104X
  [4X[32X[104X
  
  [1X18.5-6 RationalizedMat[101X
  
  [29X[2XRationalizedMat[102X( [3Xmat[103X ) [32X attribute
  
  [33X[0;0Yreturns  the  list  of  rationalized  rows of [3Xmat[103X, which must be a matrix of
  cyclotomics.  This  is  the  set of sums over orbits under the action of the
  Galois  group  of  the  entries  of  [3Xmat[103X  (see  [2XGaloisMat[102X  ([14X18.5-5[114X)), so the
  operation may be viewed as a kind of trace on the rows.[133X
  
  [33X[0;0YNote that no two rows of [3Xmat[103X should be equal.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27Xmat:= [ [ 1, 1, 1 ], [ 1, E(3), E(3)^2 ], [ 1, E(3)^2, E(3) ] ];;[127X[104X
    [4X[25Xgap>[125X [27XRationalizedMat( mat );[127X[104X
    [4X[28X[ [ 1, 1, 1 ], [ 2, -1, -1 ] ][128X[104X
  [4X[32X[104X
  
  
  [1X18.6 [33X[0;0YInternally Represented Cyclotomics[133X[101X
  
  [33X[0;0YThe  implementation  of  an  [13Xinternally represented cyclotomic[113X is based on a
  list  of  length  equal  to  its  conductor.  This  means  that the internal
  representation  of  a cyclotomic does [13Xnot[113X refer to the smallest number field
  but  the smallest [13Xcyclotomic[113X field containing it. The reason for this is the
  wish to reflect the natural embedding of two cyclotomic fields into a larger
  one  that  contains  both. With such embeddings, it is easy to construct the
  sum  or  the  product  of  two  arbitrary cyclotomics (in possibly different
  fields) as an element of a cyclotomic field.[133X
  
  [33X[0;0YThe  disadvantage  of  this approach is that the arithmetical operations are
  quite  expensive,  so  the  use of internally represented cyclotomics is not
  recommended  for  doing arithmetics over number fields, such as calculations
  with  matrices  of  cyclotomics.  But internally represented cyclotomics are
  good  enough  for  dealing  with  irrationalities  in  character tables (see
  chapter [14X71[114X).[133X
  
  [33X[0;0YFor  the  representation  of  cyclotomics  one  has  to recall that the [22Xn[122X-th
  cyclotomic  field  [22Xℚ(e_n)[122X  is  a  vector  space  of  dimension [22Xφ(n)[122X over the
  rationals where [22Xφ[122X denotes Euler's phi-function (see [2XPhi[102X ([14X15.2-2[114X)).[133X
  
  [33X[0;0YA  special  integral basis of cyclotomic fields is chosen that allows one to
  easily  convert  arbitrary sums of roots of unity into the basis, as well as
  to  convert  a  cyclotomic  represented  w.r.t. the  basis into the smallest
  possible  cyclotomic  field.  This  basis is accessible in [5XGAP[105X, see [14X60.3[114X for
  more information and references.[133X
  
  [33X[0;0YNote  that  the set of all [22Xn[122X-th roots of unity is linearly dependent for [22Xn >
  1[122X,  so  multiplication is [13Xnot[113X the multiplication of the group ring [22Xℚ⟨ e_n ⟩[122X;
  given  a  [22Xℚ[122X-basis  of  [22Xℚ(e_n)[122X  the result of the multiplication (computed as
  multiplication  of  polynomials in [22Xe_n[122X, using [22X(e_n)^n = 1[122X) will be converted
  to the basis.[133X
  
  [4X[32X  Example  [32X[104X
    [4X[25Xgap>[125X [27XE(5) * E(5)^2; ( E(5) + E(5)^4 ) * E(5)^2;[127X[104X
    [4X[28XE(5)^3[128X[104X
    [4X[28XE(5)+E(5)^3[128X[104X
    [4X[25Xgap>[125X [27X( E(5) + E(5)^4 ) * E(5);[127X[104X
    [4X[28X-E(5)-E(5)^3-E(5)^4[128X[104X
  [4X[32X[104X
  
  [33X[0;0YAn  internally  represented cyclotomic is always represented in the smallest
  cyclotomic  field  it  is  contained  in.  The  internal  coefficients  list
  coincides   with   the   external  representation  returned  by  [2XExtRepOfObj[102X
  ([14X18.1-12[114X).[133X
  
  [33X[0;0YTo  avoid calculations becoming unintentionally very long, or consuming very
  large  amounts  of  memory,  there is a limit on the conductor of internally
  represented  cyclotomics,  by default set to one million. This can be raised
  (although not lowered) using [2XSetCyclotomicsLimit[102X ([14X18.6-1[114X) and accessed using
  [2XGetCyclotomicsLimit[102X ([14X18.6-1[114X). The maximum value of the limit is [22X2^28-1[122X on [22X32[122X
  bit  systems,  and  [22X2^32[122X  on [22X64[122X bit systems. So the maximal cyclotomic field
  implemented in [5XGAP[105X is not really the field [22Xℚ^ab[122X.[133X
  
  [33X[0;0YIt  should  be emphasized that one disadvantage of representing a cyclotomic
  in  the  smallest  [13Xcyclotomic[113X  field (and not in the smallest field) is that
  arithmetic  operations  in  a  fixed  small  extension field of the rational
  number  field are comparatively expensive. For example, take a prime integer
  [22Xp[122X  and  suppose  that  we  want  to  work with a matrix group over the field
  [22Xℚ(sqrt{p})[122X.  Then  each  matrix  entry  could  be  described by two rational
  coefficients,  whereas  the  representation in the smallest cyclotomic field
  requires  [22Xp-1[122X  rational coefficients for each entry. So it is worth thinking
  about using elements in a field constructed with [2XAlgebraicExtension[102X ([14X67.1-1[114X)
  when natural embeddings of cyclotomic fields are not needed.[133X
  
  [1X18.6-1 SetCyclotomicsLimit[101X
  
  [29X[2XSetCyclotomicsLimit[102X( [3Xnewlimit[103X ) [32X function
  [29X[2XGetCyclotomicsLimit[102X(  ) [32X function
  
  [33X[0;0Y[2XGetCyclotomicsLimit[102X  returns  the  current limit on conductors of internally
  represented cyclotomic numbers[133X
  
  [33X[0;0Y[2XSetCyclotomicsLimit[102X  can  be  called  to increase the limit on conductors of
  internally  represented  cyclotomic  numbers.  Note  that computing in large
  cyclotomic   fields   using   this  representation  can  be  both  slow  and
  memory-consuming, and that other approaches may be better for some problems.
  See [14X18.6[114X.[133X
  
