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

\documentclass{article}
\usepackage{axiom}

\author{Gabriel Dos~Reis}

\begin{document}

\begin{abstract}
\end{abstract}

\tableofcontents
\eject

\section{domain Syntax}

<<domain SYNTAX Syntax>>=
import UnionType
import SetCategory
import Boolean
import Integer
import DoubleFloat
import Symbol
import SExpression
)abbrev domain SYNTAX Syntax
++ Author: Gabriel Dos Reis
++ Date Created: November 10, 2007
++ Date Last Updated: December 05, 2007
++ Description:  This domain provides a simple domain, general enough for
++ building complete representation of Spad programs as objects 
++ of a term algebra built from ground terms of type integers, foats, 
++ symbols, and strings.  
++ This domain differs from InputForm in that it represents
++ any entity in a Spad program, not just expressions.
++ Related Constructors: Boolean, Integer, Float, Symbol, String, SExpression.
++ See Also: SExpression, SetCategory.
++ The equality supported by this domain is structural.
++ Fixme: Provide direct support for boolean values, arbritrary 
++        precision float point values.
Syntax(): Public == Private where
  Public ==> Join(UnionType, SetCategory) with
    convert: % -> SExpression
      ++ convert(s) returns the s-expression representation of a syntax.

    convert: SExpression -> %
      ++ convert(s) converts an s-expression to Syntax.  Note, when `s'
      ++ is not an atom, it is expected that it designates a proper list,
      ++ e.g. a sequence of cons cells ending with nil.

    coerce: Integer -> %
      ++ coerce(i) injects the integer value `i' into the Syntax domain.
    coerce: % -> Integer
      ++ coerce(s) extracts and integer value from the syntax `s'
    autoCoerce: % -> Integer
      ++ autoCoerce(s) forcibly extracts an integer value from
      ++ the syntax `s'; no check performed.  To be called only
      ++ at the discretion of the compiler.

    coerce: DoubleFloat -> %
      ++ coerce(f) injects the float value `f' into the Syntax domain
    coerce: % -> DoubleFloat
      ++ coerce(s) extracts a float value from the syntax `s'.
    autoCoerce: % -> DoubleFloat
      ++ autoCoerce(s) forcibly extracts a float value from the syntax `s';
      ++ no check performed.  To be called only at the discretion of
      ++ the compiler

    coerce: Symbol -> %
      ++ coerce(s) injects the symbol `s' into the Syntax domain.
    coerce: % -> Symbol
      ++ coerce(s) extracts a symbol from the syntax `s'.
    autoCoerce: % -> Symbol
      ++ autoCoerce(s) forcibly extracts a symbo from the Syntax
      ++ domain `s'; no check performed.  To be called only at
      ++ at the discretion of the compiler.

    coerce: String -> %
      ++ coerce(s) injects the string value `s' into the syntax domain
    coerce: % -> String
      ++ coerce(s) extracts a string value from the syntax `s'.
    autoCoerce: % -> String
      ++ autoCoerce(s) forcibly extracts a string value from
      ++ the syntax `s'; no check performed.  To be called only at
      ++ the discretion of the compiler.

    buildSyntax: (Symbol, List %) -> %
      ++ buildSyntax(op, [a1, ..., an]) builds a syntax object 
      ++ for op(a1,...,an).
  
    buildSyntax: (%, List %) -> %
      ++ buildSyntax(op, [a1, ..., an]) builds a syntax object 
      ++ for op(a1,...,an).

    nil?: % -> Boolean
      ++ nil?(s) is true when `s' is a syntax for the constant nil.

    getOperator: % -> Union(Integer, DoubleFloat, Symbol, String, %)
      ++ getOperator(x) returns the operator, or tag, of the syntax `x'.
      ++ The value returned is itself a syntax if `x' really is an 
      ++ application of a function symbol as opposed to being an
      ++ atomic ground term.
 
    getOperands: % -> List %
      ++ getOperands(x) returns the list of operands to the operator in `x'.

    compound?: % -> Boolean
      ++ compound? x is true when `x' is not an atomic syntax.

    _case: (%, [|Integer|]) -> Boolean
      ++ x case Integer is true if `x' really is an Integer 

    _case: (%, [|DoubleFloat|]) -> Boolean
      ++ x case DoubleFloat is true if `x' really is a DoubleFloat

    _case: (%, [|Symbol|]) -> Boolean
      ++ x case Symbol is true if `x' really is a Symbol

    _case: (%, [|String|]) -> Boolean
      ++ x case String is true if `x' really is a String

  Private ==> SExpression add
    rep(x: %): SExpression ==
      x pretend SExpression

    per(x: SExpression): % ==
      x pretend %

    x = y ==
      EQUAL(x,y)$Lisp @ Boolean

    s case Integer ==
      integer? rep s

    s case DoubleFloat ==
      float? rep s

    s case String ==
      string? rep s

    s case Symbol ==
      symbol? rep s

    convert(x: %): SExpression ==
      rep x

    convert(x: SExpression): % ==
      per x

    coerce(i: Integer): % ==
      i pretend %

    autoCoerce(i: %): Integer ==                -- used for hard coercion
      i : Integer

    coerce(i: %): Integer ==
      i case Integer => i
      userError "invalid conversion target type"

    coerce(f: DoubleFloat): % ==
      f pretend %

    autoCoerce(f: %): DoubleFloat ==		-- used for hard coercion
      f : DoubleFloat

    coerce(f: %): DoubleFloat ==
      f case DoubleFloat => f
      userError "invalid conversion target type"

    coerce(s: Symbol): % ==
      s pretend %

    autoCoerce(s: %): Symbol ==                 -- used for hard coercion
      s : Symbol

    coerce(s: %): Symbol ==
      s case Symbol => s
      userError "invalid conversion target type"

    coerce(s: String): % ==
      s pretend %

    autoCoerce(s: %): String ==                 -- used for hard coercion
      s : String

    coerce(s: %): String ==
      s case String => s
      userError "invalid conversion target type"

    buildSyntax(s: Symbol, l: List %): % ==
      -- ??? ideally we should have overloaded operator `per' that convert
      -- from list of syntax to syntax.  But the compiler is at the 
      -- moment defective for non-exported overloaded operations.
      -- Furthermore, this direct call to `CONS' is currently necessary
      -- in order to have the Syntax domain compiled as early as possible
      -- in algebra boostrapping process.  It should be removed once
      -- the bootstrap process is improved.
      CONS(s,l)$Lisp @ %

    buildSyntax(op: %, l: List %): % ==
      CONS(op,l)$Lisp @ %

    nil? x ==
      null? rep x

    getOperator x ==
      atom? rep x => userError "atom as operand to getOperator"
      op := car rep x
      symbol? op => symbol op
      integer? op => integer op
      float? op => float op
      string? op => string op
      convert op
   
    compound? x ==
      pair? rep x

    getOperands x ==
      s := rep x
      atom? s => []
      [per t for t in destruct cdr s]
@


\section{domain ConstructorCall}

<<domain CTORCALL ConstructorCall>>=
import SetCategory
import Symbol
import List Syntax
)abbrev domain CTORCALL ConstructorCall
++ Author: Gabriel Dos Reis
++ Date Created: January 19, 2008
++ Date Last Updated: July 03, 2008
++ Description: This domains represents a syntax object that
++ designates a category, domain, or a package.
++ See Also: Syntax, Domain
ConstructorCall(): Public == Private where
  Public == SetCategory with
    constructorName: % -> Symbol
      ++ constructorName c returns the name of the constructor
    arguments: % -> List Syntax
      ++ arguments returns the list of syntax objects for the
      ++ arguments used to invoke the constructor.

  Private == add
    Rep == List Syntax

    constructorName x ==
      (first rep x)::Symbol

    arguments x ==
      rest rep x

    x = y ==
      rep x = rep y

    coerce x ==
      outputDomainConstructor(x)$Lisp
@

\section{The Signature domain}

<<domain SIG Signature>>=
import SetCategory
import ConstructorCall
)abbrev domain SIG Signature
++ Author: Gabriel Dos Reis
++ Date Created: January 10, 2008
++ Date Last Updated: July 14, 2008
++ Description: This is the datatype for operation signatures as
++ used by the compiler and the interpreter.
++ See also: ConstructorCall, Domain.
Signature(): Public == Private where
  Public == SetCategory with
    target: % -> ConstructorCall
      ++ target(s) returns the target type of the signature `s'.
    source: % -> List ConstructorCall
      ++ source(s) returns the list of parameter types of `s'.
  Private == add
    Rep == List ConstructorCall
    target x ==
      first rep x
    source x ==
      rest rep x
    x = y ==
      rep x = rep y
    coerce(x: %): OutputForm ==
      rarrow([s::OutputForm for s in source x]::OutputForm,
         target(x)::OutputForm)$OutputForm
@

\section{domain ElaboratedExpression}

<<domain ELABEXPR ElaboratedExpression>>=
import Boolean
import Symbol
import ConstructorCall
import List
import SExpression
)abbrev domain ELABEXPR ElaboratedExpression
++ Author: Gabriel Dos Reis
++ Date Created: January 19, 2008
++ Date Last Updated: January 20, 2008
++ Description: This domains an expresion as elaborated by the interpreter.
++ See Also: 
ElaboratedExpression(): Public == Private where
  Public ==> CoercibleTo OutputForm with
    type: % -> ConstructorCall
      ++ type(e) returns the type of the expression as computed by
      ++ the interpreter.
    constant?: % -> Boolean
      ++ constant?(e) returns true if `e' is a constant.
    getConstant: % -> Union(SExpression,"failed")
      ++ getConstant(e) retrieves the constant value of `e'e.
    variable?: % -> Boolean
      ++ variable?(e) returns true if `e' is a variable.
    getIdentifier: % -> Union(Symbol,"failed")
      ++ getIdentifier(e) retrieves the name of the variable `e'.
    callForm?: % -> Boolean
      ++ callForm?(e) is true when `e' is a call expression.
    getOperator: % -> Union(Symbol, "failed")
      ++ getOperator(e) retrieves the operator being invoked in `e',
      ++ when `e' is an expression.  
    getOperands: % -> Union(List %, "failed")
      ++ getOperands(e) returns the list of operands in `e', assuming it
      ++ is a call form.

  Private ==> add
    isAtomic(x: %): Boolean ==
      ATOM(x)$Lisp @ Boolean

    type x ==
      getMode(x)$Lisp @ ConstructorCall

    callForm? x ==
      CONSP(x)$Lisp @ Boolean

    getOperator x ==
      op: SExpression := getUnnameIfCan(x)$Lisp
      null? op => "failed"
      op pretend Symbol

    constant? x ==
      isAtomic x and 
        EQ(getUnnameIfCan(x)$Lisp, _$immediateDataSymbol$Lisp)$Lisp : Boolean

    getConstant x ==
      constant? x => getValue(x)$Lisp @ SExpression
      "failed"

    variable? x ==
      isAtomic x and not constant? x
   
    getIdentifier x ==
      variable? x => symbol (getUnname(x)$Lisp@SExpression)
      "failed"
@


\section{SpadAbstractSyntaxCategory}

<<category ASTCAT AbstractSyntaxCategory>>=
import SetCategory
import CoercibleTo Syntax
)abbrev category ASTCAT AbstractSyntaxCategory
++ Author: Gabriel Dos Reis
++ Date Created: July 5, 2008
++ Date Last Modified: July 5, 2008
++ Description: This is the category of Spad abstract syntax trees.
AbstractSyntaxCategory(): Category == 
    Join(SetCategory, CoercibleTo Syntax)
  add
    coerce(x: %): Syntax ==
      x pretend Syntax
    coerce(x: %): OutputForm ==
      x::Syntax::OutputForm 
@

\subsection{The HeadAst domain}

<<domain HEADAST HeadAst>>=
import AbstractSyntaxCategory
import Symbol
import List Symbol
)abbrev domain HEADAST HeadAst
++ Author: Gabriel Dos Reis
++ Date Created: November 10, 2007
++ Date Last Modified: July 2008
++ Description: This domain represents the header of a definition.
HeadAst(): Public == Private where
  Public == AbstractSyntaxCategory with
    headAst: List Symbol -> %
      ++ headAst [f,x1,..,xn] constructs a function definition header.
    name: % -> Symbol
      ++ name(h) returns the name of the operation defined defined.
    parameters: % -> List Symbol
      ++ parameters(h) gives the parameters specified in the
      ++ definition header `h'.
  Private == add
    Rep == List Symbol
    headAst h == per h
    name h == first rep h
    parameters h == rest rep h
@

\section{License}

<<license>>=
--Copyright (C) 2007-2008, 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>>
<<domain SYNTAX Syntax>>
<<domain CTORCALL ConstructorCall>>
<<domain ELABEXPR ElaboratedExpression>>

<<category ASTCAT AbstractSyntaxCategory>>
<<domain HEADAST HeadAst>>
@

\end{document}