path: root/src/algebra/ffcat.spad.pamphlet

\documentclass{article}
\usepackage{open-axiom}
\begin{document}
\title{\$SPAD/src/algebra ffcat.spad}
\author{Johannes Grabmeier, Alfred Scheerhorn, Barry Trager, James Davenport}
\maketitle
\begin{abstract}
\end{abstract}
\eject
\tableofcontents
\eject
\begin{verbatim}
-- 28.01.93: AS and JG:another Error in discreteLog(.,.) in FFIEDLC corrected.
-- 08.05.92: AS  Error in discreteLog(.,.) in FFIEDLC corrected.
-- 03.04.92: AS  Barry Trager added package FFSLPE and some functions to FFIELDC
-- 25.02.92: AS  added following functions in FAXF: impl.of mrepresents,
--               linearAssociatedExp,linearAssociatedLog, linearAssociatedOrder
-- 18.02.92: AS: more efficient version of degree added,
--               first version of degree in FAXF set into comments
-- 18.06.91: AS: general version of minimalPolynomial added
-- 08.05.91: JG, AS implementation of missing functions in FFC and FAXF
-- 04.05.91: JG: comments
-- 04.04.91: JG: old version of charthRoot in FFC was dropped

-- Fields with finite characteristic
\end{verbatim}
\section{category FPC FieldOfPrimeCharacteristic}
<<category FPC FieldOfPrimeCharacteristic>>=
)abbrev category FPC FieldOfPrimeCharacteristic
++ Author: J. Grabmeier, A. Scheerhorn
++ Date Created: 10 March 1991
++ Date Last Updated: 31 March 1991
++ Basic Operations: _+, _*
++ Related Constructors:
++ Also See:
++ AMS Classifications:
++ Keywords: field, finite field, prime characteristic
++ References:
++  J. Grabmeier, A. Scheerhorn: Finite Fields in AXIOM.
++  AXIOM Technical Report Series, ATR/5 NP2522.
++ Description:
++  FieldOfPrimeCharacteristic is the category of fields of prime
++  characteristic, e.g. finite fields, algebraic closures of
++  fields of prime characteristic, transcendental extensions of
++  of fields of prime characteristic.
FieldOfPrimeCharacteristic:Category == _
  Join(Field,CharacteristicNonZero) with
    order: $ -> OnePointCompletion PositiveInteger
      ++ order(a) computes the order of an element in the multiplicative
      ++ group of the field.
      ++ Error: if \spad{a} is 0.
    discreteLog: ($,$) -> Union(NonNegativeInteger,"failed")
      ++ discreteLog(b,a) computes s with \spad{b**s = a} if such an s exists.
    primeFrobenius: $ -> $
      ++ primeFrobenius(a) returns \spad{a ** p} where p is the characteristic.
    primeFrobenius: ($,NonNegativeInteger) -> $
      ++ primeFrobenius(a,s) returns \spad{a**(p**s)} where p
      ++ is the characteristic.
  add
    primeFrobenius(a) == a ** characteristic$%
    primeFrobenius(a,s) == a ** (characteristic$%**s)

@
\section{category XF ExtensionField}
<<category XF ExtensionField>>=
)abbrev category XF ExtensionField
++ Author: J. Grabmeier, A. Scheerhorn
++ Date Created: 10 March 1991
++ Date Last Updated: 31 March 1991
++ Basic Operations: _+, _*, extensionDegree, algebraic?, transcendent?
++ Related Constructors:
++ Also See:
++ AMS Classifications:
++ Keywords: field, extension field
++ References:
++  J. Grabmeier, A. Scheerhorn: Finite Fields in AXIOM.
++  AXIOM Technical Report Series, ATR/5 NP2522.
++ Description:
++  ExtensionField {\em F} is the category of fields which extend
++  the field F
ExtensionField(F:Field) : Category  == Join(Field,RetractableTo F,VectorSpace F) with
    if F has CharacteristicZero then CharacteristicZero
    if F has CharacteristicNonZero then FieldOfPrimeCharacteristic
    algebraic? : $ -> Boolean
      ++ algebraic?(a) tests whether an element \spad{a} is algebraic with
      ++ respect to the ground field F.
    transcendent? : $ -> Boolean
      ++ transcendent?(a) tests whether an element \spad{a} is transcendent
      ++ with respect to the ground field F.
    inGroundField?: $ -> Boolean
      ++ inGroundField?(a) tests whether an element \spad{a}
      ++ is already in the ground field F.
    degree : $ -> OnePointCompletion PositiveInteger
      ++ degree(a) returns the degree of minimal polynomial of an element
      ++ \spad{a} if \spad{a} is algebraic
      ++ with respect to the ground field F, and \spad{infinity} otherwise.
    extensionDegree : () -> OnePointCompletion PositiveInteger
      ++ extensionDegree() returns the degree of the field extension if the
      ++ extension is algebraic, and \spad{infinity} if it is not.
    transcendenceDegree : () -> NonNegativeInteger
      ++ transcendenceDegree() returns the transcendence degree of the
      ++ field extension, 0 if the extension is algebraic.
    -- perhaps more absolute degree functions
    if F has Finite then
      FieldOfPrimeCharacteristic
      Frobenius: $ -> $
        ++ Frobenius(a) returns \spad{a ** q} where q is the \spad{size()$F}.
      Frobenius:   ($,NonNegativeInteger) -> $
        ++ Frobenius(a,s) returns \spad{a**(q**s)} where q is the size()$F.
  add
    algebraic?(a) == not infinite? (degree(a)@OnePointCompletion_
      (PositiveInteger))$OnePointCompletion(PositiveInteger)
    transcendent? a == infinite?(degree(a)@OnePointCompletion _
      (PositiveInteger))$OnePointCompletion(PositiveInteger)
    if F has Finite then
      Frobenius(a) == a ** size()$F
      Frobenius(a,s) == a ** (size()$F ** s)

@

\section{category FAXF FiniteAlgebraicExtensionField}

<<category FAXF FiniteAlgebraicExtensionField>>=
import Boolean
import NonNegativeInteger
import PositiveInteger
import Vector
import Matrix
import SparseUnivariatePolynomial
import OnePointCompletion
import CardinalNumber
)abbrev category FAXF FiniteAlgebraicExtensionField
++ Author: J. Grabmeier, A. Scheerhorn
++ Date Created: 11 March 1991
++ Date Last Updated: 31 March 1991
++ Basic Operations: _+, _*, extensionDegree,
++ Related Constructors:
++ Also See:
++ AMS Classifications:
++ Keywords: field, extension field, algebraic extension, finite extension
++ References:
++  R.Lidl, H.Niederreiter: Finite Field, Encycoldia of Mathematics and
++  Its Applications, Vol. 20, Cambridge Univ. Press, 1983, ISBN 0 521 30240 4
++  J. Grabmeier, A. Scheerhorn: Finite Fields in AXIOM.
++  AXIOM Technical Report Series, ATR/5 NP2522.
++ Description:
++  FiniteAlgebraicExtensionField {\em F} is the category of fields
++  which are finite algebraic extensions of the field {\em F}.
++  If {\em F} is finite then any finite algebraic extension of {\em F} is finite, too.
++  Let {\em K} be a finite algebraic extension of the finite field {\em F}.
++  The exponentiation of elements of {\em K} defines a Z-module structure
++  on the multiplicative group of {\em K}. The additive group of {\em K}
++  becomes a module over the ring of polynomials over {\em F} via the operation
++  \spadfun{linearAssociatedExp}(a:K,f:SparseUnivariatePolynomial F)
++  which is linear over {\em F}, i.e. for elements {\em a} from {\em K},
++  {\em c,d} from {\em F} and {\em f,g} univariate polynomials over {\em F}
++  we have \spadfun{linearAssociatedExp}(a,cf+dg) equals {\em c} times
++  \spadfun{linearAssociatedExp}(a,f) plus {\em d} times
++  \spadfun{linearAssociatedExp}(a,g).
++  Therefore \spadfun{linearAssociatedExp} is defined completely by
++  its action on  monomials from {\em F[X]}:
++  \spadfun{linearAssociatedExp}(a,monomial(1,k)\$SUP(F)) is defined to be
++  \spadfun{Frobenius}(a,k) which is {\em a**(q**k)} where {\em q=size()\$F}.
++  The operations order and discreteLog associated with the multiplicative
++  exponentiation have additive analogues associated to the operation
++  \spadfun{linearAssociatedExp}. These are the functions
++  \spadfun{linearAssociatedOrder} and \spadfun{linearAssociatedLog},
++  respectively.

FiniteAlgebraicExtensionField(F : Field) : Category == _
  Join(ExtensionField F, RetractableTo F) with
  -- should be unified with algebras
  -- Join(ExtensionField F, FramedAlgebra F, RetractableTo F) with
    basis : () -> Vector $
      ++ basis() returns a fixed basis of \$ as \spad{F}-vectorspace.
    basis : PositiveInteger -> Vector $
      ++ basis(n) returns a fixed basis of a subfield of \$ as
      ++ \spad{F}-vectorspace.
    coordinates : $ -> Vector F
      ++ coordinates(a) returns the coordinates of \spad{a} with respect
      ++ to the fixed \spad{F}-vectorspace basis.
    coordinates : Vector $ -> Matrix F
      ++ coordinates([v1,...,vm]) returns the coordinates of the
      ++ vi's with to the fixed basis.  The coordinates of vi are
      ++ contained in the ith row of the matrix returned by this
      ++ function.
    represents:  Vector F -> $
      ++ represents([a1,..,an]) returns \spad{a1*v1 + ... + an*vn}, where
      ++ v1,...,vn are the elements of the fixed basis.
    minimalPolynomial: $ -> SparseUnivariatePolynomial F
      ++ minimalPolynomial(a) returns the minimal polynomial of an
      ++ element \spad{a} over the ground field F.
    definingPolynomial: () -> SparseUnivariatePolynomial F
      ++ definingPolynomial() returns the polynomial used to define
      ++ the field extension.
    extensionDegree : () ->  PositiveInteger
      ++ extensionDegree() returns the degree of field extension.
    degree : $ -> PositiveInteger
      ++ degree(a) returns the degree of the minimal polynomial of an
      ++ element \spad{a} over the ground field F.
    norm: $  -> F
      ++ norm(a) computes the norm of \spad{a} with respect to the
      ++ field considered as an algebra with 1 over the ground field F.
    trace: $ -> F
      ++ trace(a) computes the trace of \spad{a} with respect to
      ++ the field considered as an algebra with 1 over the ground field F.
    if F has Finite then
      FiniteFieldCategory
      minimalPolynomial: ($,PositiveInteger) -> SparseUnivariatePolynomial $
        ++ minimalPolynomial(x,n) computes the minimal polynomial of x over
        ++ the field of extension degree n over the ground field F.
      norm: ($,PositiveInteger)  -> $
        ++ norm(a,d) computes the norm of \spad{a} with respect to the field of
        ++ extension degree d over the ground field of size.
        ++ Error: if d does not divide the extension degree of \spad{a}.
        ++ Note: norm(a,d) = reduce(*,[a**(q**(d*i)) for i in 0..n/d])
      trace: ($,PositiveInteger)   -> $
        ++ trace(a,d) computes the trace of \spad{a} with respect to the
        ++ field of extension degree d over the ground field of size q.
        ++ Error: if d does not divide the extension degree of \spad{a}.
        ++ Note: \spad{trace(a,d) = reduce(+,[a**(q**(d*i)) for i in 0..n/d])}.
      createNormalElement: () -> $
        ++ createNormalElement() computes a normal element over the ground
        ++ field F, that is,
        ++ \spad{a**(q**i), 0 <= i < extensionDegree()} is an F-basis,
        ++ where \spad{q = size()\$F}.
        ++ Reference: Such an element exists Lidl/Niederreiter: Theorem 2.35.
      normalElement: () -> $
        ++ normalElement() returns a element, normal over the ground field F,
        ++ i.e. \spad{a**(q**i), 0 <= i < extensionDegree()} is an F-basis,
        ++ where \spad{q = size()\$F}.
        ++ At the first call, the element is computed by
        ++ \spadfunFrom{createNormalElement}{FiniteAlgebraicExtensionField}
        ++ then cached in a global variable.
        ++ On subsequent calls, the element is retrieved by referencing the
        ++ global variable.
      normal?: $ -> Boolean
        ++ normal?(a) tests whether the element \spad{a} is normal over the
        ++ ground field F, i.e.
        ++ \spad{a**(q**i), 0 <= i <= extensionDegree()-1} is an F-basis,
        ++ where \spad{q = size()\$F}.
        ++ Implementation according to Lidl/Niederreiter: Theorem 2.39.
      generator: () -> $
        ++ generator() returns a root of the defining polynomial.
        ++ This element generates the field as an algebra over the ground field.
      linearAssociatedExp:($,SparseUnivariatePolynomial F) -> $
        ++ linearAssociatedExp(a,f) is linear over {\em F}, i.e.
        ++ for elements {\em a} from {\em \$}, {\em c,d} form {\em F} and
        ++ {\em f,g} univariate polynomials over {\em F} we have
        ++ \spadfun{linearAssociatedExp}(a,cf+dg) equals {\em c} times
        ++ \spadfun{linearAssociatedExp}(a,f) plus {\em d} times
        ++ \spadfun{linearAssociatedExp}(a,g). Therefore
        ++ \spadfun{linearAssociatedExp} is defined completely by its action on
        ++ monomials from {\em F[X]}:
        ++ \spadfun{linearAssociatedExp}(a,monomial(1,k)\$SUP(F)) is defined to
        ++ be \spadfun{Frobenius}(a,k) which is {\em a**(q**k)},
        ++ where {\em q=size()\$F}.
      linearAssociatedOrder: $ -> SparseUnivariatePolynomial F
        ++ linearAssociatedOrder(a) retruns the monic polynomial {\em g} of
        ++ least degree, such that \spadfun{linearAssociatedExp}(a,g) is 0.
      linearAssociatedLog: $ -> SparseUnivariatePolynomial F
        ++ linearAssociatedLog(a) returns a polynomial {\em g}, such that
        ++ \spadfun{linearAssociatedExp}(normalElement(),g) equals {\em a}.
      linearAssociatedLog: ($,$) -> Union(SparseUnivariatePolynomial F,"failed")
        ++ linearAssociatedLog(b,a) returns a polynomial {\em g}, such that the
        ++ \spadfun{linearAssociatedExp}(b,g) equals {\em a}.
        ++ If there is no such polynomial {\em g}, then
        ++ \spadfun{linearAssociatedLog} fails.
  add
    I   ==> Integer
    PI  ==> PositiveInteger
    NNI ==> NonNegativeInteger
    SUP ==> SparseUnivariatePolynomial
    DLP ==> DiscreteLogarithmPackage

    represents(v) ==
      a:$:=0
      b:=basis()
      for i in 1..extensionDegree()@PI repeat
        a:=a+(v.i)*(b.i)
      a
    transcendenceDegree() == 0$NNI
    dimension() == (#basis()) ::NonNegativeInteger::CardinalNumber
    extensionDegree():OnePointCompletion(PositiveInteger) ==
      (#basis()) :: PositiveInteger::OnePointCompletion(PositiveInteger)
    degree(a):OnePointCompletion(PositiveInteger) ==
      degree(a)@PI::OnePointCompletion(PositiveInteger)

    coordinates(v:Vector $) ==
      m := new(#v, extensionDegree(), 0)$Matrix(F)
      for i in minIndex v .. maxIndex v for j in minRowIndex m .. repeat
        setRow!(m, j, coordinates qelt(v, i))
      m
    algebraic? a == true
    transcendent? a == false
    extensionDegree(): PositiveInteger == (#basis()) :: PositiveInteger
    -- degree a == degree(minimalPolynomial a)$SUP(F) :: PI
    trace a ==
      b := basis()
      abs : F := 0
      for i in 1..#b repeat
        abs := abs + coordinates(a*b.i).i
      abs
    norm a ==
      b := basis()
      m := new(#b,#b, 0)$Matrix(F)
      for i in 1..#b repeat
        setRow!(m,i, coordinates(a*b.i))
      determinant(m)
    if F has Finite then
      linearAssociatedExp(x,f) ==
        erg:$:=0
        y:=x
        for i in 0..degree(f) repeat
          erg:=erg + coefficient(f,i) * y
          y:=Frobenius(y)
        erg
      linearAssociatedLog(b,x) ==
        x=0 => 0
        l:List List F:=[entries coordinates b]
        a:$:=b
        extdeg:NNI:=extensionDegree()@PI
        for i in 2..extdeg repeat
          a:=Frobenius(a)
          l:=concat(l,entries coordinates a)$(List List F)
        l:=concat(l,entries coordinates x)$(List List F)
        m1:=rowEchelon transpose matrix(l)$(Matrix F)
        v:=zero(extdeg)$(Vector F)
        rown:I:=1
        for i in 1..extdeg repeat
          if qelt(m1,rown,i) = 1$F then
            v.i:=qelt(m1,rown,extdeg+1)
            rown:=rown+1
        p:=+/[monomial(v.(i+1),i::NNI) for i in 0..(#v-1)]
        p=0 =>
         messagePrint("linearAssociatedLog: second argument not in_
                       group generated by first argument")$OutputForm
         "failed"
        p
      linearAssociatedLog(x) == linearAssociatedLog(normalElement(),x) ::
                              SparseUnivariatePolynomial(F)
      linearAssociatedOrder(x) ==
        x=0 => 0
        l:List List F:=[entries coordinates x]
        a:$:=x
        for i in 1..extensionDegree()@PI repeat
          a:=Frobenius(a)
          l:=concat(l,entries coordinates a)$(List List F)
        v:=first nullSpace transpose matrix(l)$(Matrix F)
        +/[monomial(v.(i+1),i::NNI) for i in 0..(#v-1)]

      charthRoot(x):Union($,"failed") ==
        (charthRoot(x)@$)::Union($,"failed")
      -- norm(e) == norm(e,1) pretend F
      -- trace(e) == trace(e,1) pretend F
      minimalPolynomial(a,n) ==
        extensionDegree()@PI rem n ~= 0 =>
          error "minimalPolynomial: 2. argument must divide extension degree"
        f:SUP $:=monomial(1,1)$(SUP $) - monomial(a,0)$(SUP $)
        u:$:=Frobenius(a,n)
        while not(u = a) repeat
          f:=f * (monomial(1,1)$(SUP $) - monomial(u,0)$(SUP $))
          u:=Frobenius(u,n)
        f
      norm(e,s) ==
        qr := divide(extensionDegree(), s)
        zero?(qr.remainder) =>
          pow := (size()-1) quo (size()$F ** s - 1)
          e ** (pow::NonNegativeInteger)
        error "norm: second argument must divide degree of extension"
      trace(e,s) ==
        qr:=divide(extensionDegree(),s)
        q:=size()$F
        zero?(qr.remainder) =>
          a:$:=0
          for i in 0..qr.quotient-1 repeat
            a:=a + e**(q**(s*i))
          a
        error "trace: second argument must divide degree of extension"
      size() == size()$F ** extensionDegree()
      createNormalElement() ==
        characteristic$% = size() => 1
        res : $
        for i in 1.. repeat
          res := index(i :: PI)
          not inGroundField? res =>
            normal? res => return res
        -- theorem: there exists a normal element, this theorem is
        -- unknown to the compiler
        res
      normal?(x:$) ==
        p:SUP $:=(monomial(1,extensionDegree()) - monomial(1,0))@(SUP $)
        f:SUP $:= +/[monomial(Frobenius(x,i),i)$(SUP $) _
                   for i in 0..extensionDegree()-1]
        gcd(p,f) = 1 => true
        false
      degree(a: %): PositiveInteger ==
        y:$:=Frobenius a
        deg:PI:=1
        while y~=a repeat
          y := Frobenius(y)
          deg:=deg+1
        deg

@
\section{package DLP DiscreteLogarithmPackage}
<<package DLP DiscreteLogarithmPackage>>=
)abbrev package DLP DiscreteLogarithmPackage
++ Author: J. Grabmeier, A. Scheerhorn
++ Date Created: 12 March 1991
++ Date Last Updated: 31 March 1991
++ Basic Operations:
++ Related Constructors:
++ Also See:
++ AMS Classifications:
++ Keywords: discrete logarithm
++ References:
++  J. Grabmeier, A. Scheerhorn: Finite Fields in AXIOM.
++  AXIOM Technical Report Series, ATR/5 NP2522.
++ Description:
++  DiscreteLogarithmPackage implements help functions for discrete logarithms
++  in monoids using small cyclic groups.

DiscreteLogarithmPackage(M): public == private where
  M : Join(Monoid,Finite) with
   **: (M,Integer) -> M
	++ x ** n returns x raised to the integer power n
  public ==> with
    shanksDiscLogAlgorithm:(M,M,NonNegativeInteger)->  _
        Union(NonNegativeInteger,"failed")
      ++ shanksDiscLogAlgorithm(b,a,p) computes s with \spad{b**s = a} for
      ++ assuming that \spad{a} and b are elements in a 'small' cyclic group of
      ++ order p by Shank's algorithm.
      ++ Note: this is a subroutine of the function \spadfun{discreteLog}.
  I   ==> Integer
  PI  ==> PositiveInteger
  NNI ==> NonNegativeInteger
  SUP ==> SparseUnivariatePolynomial
  DLP ==> DiscreteLogarithmPackage

  private ==> add
    shanksDiscLogAlgorithm(logbase,c,p) ==
      limit:Integer:= 30
      -- for logarithms up to cyclic groups of order limit a full
      -- logarithm table is computed
      p < limit =>
        a:M:=1
        disclog:Integer:=0
        found:Boolean:=false
        for i in 0..p-1 while not found repeat
          a = c =>
            disclog:=i
            found:=true
          a:=a*logbase
        not found =>
          messagePrint("discreteLog: second argument not in cyclic group_
 generated by first argument")$OutputForm
          "failed"
        disclog pretend NonNegativeInteger
      l:Integer:=length(p)$Integer
      if odd?(l)$Integer then n:Integer:= shift(p,-(l quo 2))
                         else n:Integer:= shift(1,(l quo 2))
      a:M:=1
      exptable : Table(PI,NNI) :=table()$Table(PI,NNI)
      for i in (0::NNI)..(n-1)::NNI repeat
        insert!([lookup(a),i::NNI]$Record(key:PI,entry:NNI),_
                  exptable)$Table(PI,NNI)
        a:=a*logbase
      found := false
      end := (p-1) quo n
      disclog:Integer:=0
      a := c
      b := logbase ** (-n)
      for i in 0..end while not found repeat
        rho:= search(lookup(a),exptable)_
              $Table(PositiveInteger,NNI)
        rho case NNI =>
          found := true
          disclog:= n * i + rho pretend Integer
        a := a * b
      not found =>
        messagePrint("discreteLog: second argument not in cyclic group_
 generated by first argument")$OutputForm
        "failed"
      disclog pretend NonNegativeInteger

@

\section{category FFIELDC FiniteFieldCategory}

<<category FFIELDC FiniteFieldCategory>>=
import Boolean
import Integer
import NonNegativeInteger
import PositiveInteger
import Matrix
import List
import Table
import OnePointCompletion
import SparseUnivariatePolynomial
)abbrev category FFIELDC FiniteFieldCategory
++ Author: J. Grabmeier, A. Scheerhorn
++ Date Created: 11 March 1991
++ Date Last Updated: 31 March 1991
++ Basic Operations: _+, _*, extensionDegree, order, primitiveElement
++ Related Constructors:
++ Also See:
++ AMS Classifications:
++ Keywords: field, extension field, algebraic extension, finite field
++  Galois field
++ References:
++  D.Lipson, Elements of Algebra and Algebraic Computing, The
++  Benjamin/Cummings Publishing Company, Inc.-Menlo Park, California, 1981.
++  J. Grabmeier, A. Scheerhorn: Finite Fields in AXIOM.
++  AXIOM Technical Report Series, ATR/5 NP2522.
++ Description:
++  FiniteFieldCategory is the category of finite fields

FiniteFieldCategory() : Category ==_
  Join(FieldOfPrimeCharacteristic,Finite,StepThrough,DifferentialRing) with
--                 ,PolynomialFactorizationExplicit) with
    charthRoot: $ -> $
      ++ charthRoot(a) takes the characteristic'th root of {\em a}.
      ++ Note: such a root is alway defined in finite fields.
    conditionP: Matrix $ -> Union(Vector $,"failed")
      ++ conditionP(mat), given a matrix representing a homogeneous system
      ++ of equations, returns a vector whose characteristic'th powers
      ++ is a non-trivial solution, or "failed" if no such vector exists.
    -- the reason for implementing the following function is that we
    -- can implement the functions order, getGenerator and primitive? on
    -- category level without computing the, may be time intensive,
    -- factorization of size()-1 at every function call again.
    factorsOfCyclicGroupSize:_
      () -> List Record(factor:Integer,exponent:Integer)
      ++ factorsOfCyclicGroupSize() returns the factorization of size()-1
    -- the reason for implementing the function tableForDiscreteLogarithm
    -- is that we can implement the functions discreteLog and
    -- shanksDiscLogAlgorithm on category level
    -- computing the necessary exponentiation tables in the respective
    -- domains once and for all
    -- absoluteDegree : $ -> PositiveInteger
    --  ++ degree of minimal polynomial, if algebraic with respect
    --  ++ to the prime subfield
    tableForDiscreteLogarithm: Integer -> _
             Table(PositiveInteger,NonNegativeInteger)
      ++ tableForDiscreteLogarithm(a,n) returns a table of the discrete
      ++ logarithms of \spad{a**0} up to \spad{a**(n-1)} which, called with
      ++ key \spad{lookup(a**i)} returns i for i in \spad{0..n-1}.
      ++ Error: if not called for prime divisors of order of
      ++        multiplicative group.
    createPrimitiveElement: () -> $
      ++ createPrimitiveElement() computes a generator of the (cyclic)
      ++ multiplicative group of the field.
      -- RDJ: Are these next lines to be included?
      -- we run through the field and test, algorithms which construct
      -- elements of larger order were found to be too slow
    primitiveElement: () -> $
      ++ primitiveElement() returns a primitive element stored in a global
      ++ variable in the domain.
      ++ At first call, the primitive element is computed
      ++ by calling \spadfun{createPrimitiveElement}.
    primitive?: $ -> Boolean
      ++ primitive?(b) tests whether the element b is a generator of the
      ++ (cyclic) multiplicative group of the field, i.e. is a primitive
      ++ element.
      ++ Implementation Note: see ch.IX.1.3, th.2 in D. Lipson.
    discreteLog: $ -> NonNegativeInteger
      ++ discreteLog(a) computes the discrete logarithm of \spad{a}
      ++ with respect to \spad{primitiveElement()} of the field.
    order: $ -> PositiveInteger
      ++ order(b) computes the order of an element b in the multiplicative
      ++ group of the field.
      ++ Error: if b equals 0.
    representationType: () -> Union("prime","polynomial","normal","cyclic")
      ++ representationType() returns the type of the representation, one of:
      ++ \spad{prime}, \spad{polynomial}, \spad{normal}, or \spad{cyclic}.
  add
    I   ==> Integer
    PI  ==> PositiveInteger
    NNI ==> NonNegativeInteger
    SUP ==> SparseUnivariatePolynomial
    DLP ==> DiscreteLogarithmPackage

    -- exported functions

    differentiate x          == 0
    init() == 0
    nextItem a ==
      a := index(lookup(a)+1)
      zero? a => nothing
      just a
    order(e):OnePointCompletion(PositiveInteger) ==
      (order(e)@PI)::OnePointCompletion(PositiveInteger)

    conditionP(mat:Matrix $) ==
      l:=nullSpace mat
      empty? l or every?(zero?, first l) => "failed"
      map(charthRoot,first l)
    charthRoot(x:$):$ == x**(size()$% quo characteristic$%)
    charthRoot(x:%):Union($,"failed") ==
        (charthRoot(x)@$)::Union($,"failed")
    createPrimitiveElement() ==
      sm1  : PositiveInteger := (size()$%-1) pretend PositiveInteger
      start : Integer :=
        -- in the polynomial case, index from 1 to characteristic-1
        -- gives prime field elements
        representationType() = "polynomial" => characteristic$%
        1
      found : Boolean := false
      e : $
      for i in start..  while not found repeat
        e := index(i::PositiveInteger)
        found := (order(e) = sm1)
      e
    primitive? a ==
      -- add special implementation for prime field case
      zero?(a) => false
      explist := factorsOfCyclicGroupSize()
      q:=(size()$%-1)@Integer
      equalone : Boolean := false
      for exp in explist while not equalone repeat
        equalone := one?(a**(q quo exp.factor))
      not equalone
    order(e: %): PositiveInteger ==
      e = 0 => error "order(0) is not defined "
      ord:Integer:= size()$%-1 -- order e divides ord
      lof:=factorsOfCyclicGroupSize()
      for rec in lof repeat -- run through prime divisors
        a := ord quo (primeDivisor := rec.factor)
        goon := one?(e**a)
        -- run through exponents of the prime divisors
        for j in 0..(rec.exponent)-2 while goon repeat
          -- as long as we get (e**ord = 1) we
          -- continue dividing by primeDivisor
          ord := a
          a := ord quo primeDivisor
          goon := one?(e**a)
        if goon then ord := a
        -- as we do a top down search we have found the
        -- correct exponent of primeDivisor in order e
        -- and continue with next prime divisor
      ord pretend PositiveInteger
    discreteLog(b) ==
      zero?(b) => error "discreteLog: logarithm of zero"
      faclist:=factorsOfCyclicGroupSize()
      a:=b
      gen:=primitiveElement()
      -- in GF(2) its necessary to have discreteLog(1) = 1
      b = gen => 1
      disclog:Integer:=0
      mult:Integer:=1
      groupord := (size()$% - 1)@Integer
      exp:Integer:=groupord
      for f in faclist repeat
        fac:=f.factor
        for t in 0..f.exponent-1 repeat
          exp:=exp quo fac
          -- shanks discrete logarithm algorithm
          exptable:=tableForDiscreteLogarithm(fac)
          n:=#exptable
          c:=a**exp
          end:=(fac - 1) quo n
          found:=false
          disc1:Integer:=0
          for i in 0..end while not found repeat
            rho:= search(lookup(c),exptable)_
                  $Table(PositiveInteger,NNI)
            rho case NNI =>
              found := true
              disc1:=((n * i + rho)@Integer) * mult
            c:=c* gen**((groupord quo fac) * (-n))
          not found => error "discreteLog: ?? discrete logarithm"
          -- end of shanks discrete logarithm algorithm
          mult := mult * fac
          disclog:=disclog+disc1
          a:=a * (gen ** (-disc1))
      disclog pretend NonNegativeInteger

    discreteLog(logbase,b) ==
      zero?(b) =>
        messagePrint("discreteLog: logarithm of zero")$OutputForm
        "failed"
      zero?(logbase) =>
        messagePrint("discreteLog: logarithm to base zero")$OutputForm
        "failed"
      b = logbase => 1
      not zero?((groupord:=order(logbase)@PI) rem order(b)@PI) =>
         messagePrint("discreteLog: second argument not in cyclic group _
generated by first argument")$OutputForm
         "failed"
      faclist:=factors factor groupord
      a:=b
      disclog:Integer:=0
      mult:Integer:=1
      exp:Integer:= groupord
      for f in faclist repeat
        fac:=f.factor
        primroot:= logbase ** (groupord quo fac)
        for t in 0..f.exponent-1 repeat
          exp:=exp quo fac
          rhoHelp:= shanksDiscLogAlgorithm(primroot,_
                a**exp,fac pretend NonNegativeInteger)$DLP($)
          rhoHelp case "failed" => return "failed"
          rho := (rhoHelp :: NNI) * mult
          disclog := disclog + rho
          mult := mult * fac
          a:=a * (logbase ** (-rho))
      disclog pretend NonNegativeInteger

    FP ==> SparseUnivariatePolynomial($)
    FRP ==> Factored FP
    f,g:FP
    squareFreePolynomial(f:FP):FRP ==
          squareFree(f)$UnivariatePolynomialSquareFree($,FP)
    factorPolynomial(f:FP):FRP == factor(f)$DistinctDegreeFactorize($,FP)
    factorSquareFreePolynomial(f:FP):FRP ==
        f = 0 => 0
        flist := distdfact(f,true)$DistinctDegreeFactorize($,FP)
        (flist.cont :: FP) *
            (*/[primeFactor(u.irr,u.pow) for u in flist.factors])
    gcdPolynomial(f:FP,g:FP):FP ==
         gcd(f,g)$EuclideanDomain_&(FP)

@

\section{package FFSLPE FiniteFieldSolveLinearPolynomialEquation}
<<package FFSLPE FiniteFieldSolveLinearPolynomialEquation>>=
)abbrev package FFSLPE FiniteFieldSolveLinearPolynomialEquation
++ Author: Davenport
++ Date Created: 1991
++ Date Last Updated:
++ Basic Functions:
++ Related Constructors:
++ Also See:
++ AMS Classifications:
++ Keywords:
++ References:
++ Description:
++ This package solves linear diophantine equations for Bivariate polynomials
++ over finite fields

FiniteFieldSolveLinearPolynomialEquation(F:FiniteFieldCategory,
                                        FP:UnivariatePolynomialCategory F,
                                        FPP:UnivariatePolynomialCategory FP): with
   solveLinearPolynomialEquation: (List FPP, FPP) -> Union(List FPP,"failed")
              ++ solveLinearPolynomialEquation([f1, ..., fn], g)
              ++ (where the fi are relatively prime to each other)
              ++ returns a list of ai such that
              ++ \spad{g/prod fi = sum ai/fi}
              ++ or returns "failed" if no such list of ai's exists.
  == add
     oldlp:List FPP := []
     slpePrime: FP := monomial(1,1)
     oldtable:Vector List FPP := []
     lp: List FPP
     p: FPP
     import DistinctDegreeFactorize(F,FP)
     solveLinearPolynomialEquation(lp,p) ==
       if (oldlp ~= lp) then
          -- we have to generate a new table
          deg:= +/[degree u for u in lp]
          ans:Union(Vector List FPP,"failed"):="failed"
          slpePrime:=monomial(1,1)+monomial(1,0)   -- x+1: our starting guess
          while (ans case "failed") repeat
            ans:=tablePow(deg,slpePrime,lp)$GenExEuclid(FP,FPP)
            if (ans case "failed") then
               slpePrime:= nextItem(slpePrime)::FP
               while (degree slpePrime > 1) and
                     not irreducible? slpePrime repeat
                 slpePrime := nextItem(slpePrime)::FP
          oldtable:=(ans:: Vector List FPP)
       answer:=solveid(p,slpePrime,oldtable)
       answer

@
\section{License}
<<license>>=
--Copyright (c) 1991-2002, The Numerical ALgorithms Group Ltd.
--All rights reserved.
--Copyright (C) 2007-2010, Gabriel Dos Reis.
--All rights reserved.
--
--Redistribution and use in source and binary forms, with or without
--modification, are permitted provided that the following conditions are
--met:
--
--    - Redistributions of source code must retain the above copyright
--      notice, this list of conditions and the following disclaimer.
--
--    - Redistributions in binary form must reproduce the above copyright
--      notice, this list of conditions and the following disclaimer in
--      the documentation and/or other materials provided with the
--      distribution.
--
--    - Neither the name of The Numerical ALgorithms Group Ltd. nor the
--      names of its contributors may be used to endorse or promote products
--      derived from this software without specific prior written permission.
--
--THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
--IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
--TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
--PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
--OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
--EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
--PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
--PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
--LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
--NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
--SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
@
<<*>>=
<<license>>

<<package DLP DiscreteLogarithmPackage>>
<<category FPC FieldOfPrimeCharacteristic>>
<<category XF ExtensionField>>
<<category FAXF FiniteAlgebraicExtensionField>>
<<category FFIELDC FiniteFieldCategory>>
<<package FFSLPE FiniteFieldSolveLinearPolynomialEquation>>
@
\eject
\begin{thebibliography}{99}
\bibitem{1} nothing
\end{thebibliography}
\end{document}