README.html

<title>Extensible Haskell front-end for the Programatica projects -
        directory base/parse2</title>

<h1>Extensible Haskell front-end for the Programatica projects</h1>

<h2>What is in this directory (base/defs)</h2>

<p>
This directory defines functions for identifiers in abstract syntax trees.
We use overloading to allow the same function names to be reused in extended
versions of the syntax. The functions are thus methods in a small number of
classes. These classes are:

<table>
<tr><th>Class
    <th>Description
<tr><td><b class=tcon>DefinedNames</b>
        <br><b class=tcon>ClassMethods</b>
	<br>...
    <td>A method to compute the set of names introduced by declarations and
        patterns. Also, in a separate class, a method to compute the set of
	methods defined by a class declaration.
<tr><td><b class=tcon>FreeNames
    <td>A method to compute the set of names that occur free in declarations,
        expressions, patterns and type expressions
<tr><td><b class=tcon>NameMaps</b>
    <td>Traverse an abstract syntax tree and apply a given function to every
        identifier. Some context information is provided by the
	traversal function.
<tr><td><b class=tcon>ReAssoc</b>
    <td>Rewrite expressions according to operator precedence and associativity.
<tr><td><b class=tcon>ScopeNames</b>
    <td>Pair identifiers with the environment in which it should be interpreted.
</table>

<p>
For each of these classes, there is
<ol>
<li>one module defining the class and some related auxiliary functions,
<li>one module defining re-usable instances for the base language structure,
<li>one module defining the knot-tying recursive definitions for the
    base language (Haskell 98).
</ol>

<table>
<caption>Modules</caption>
<tr><th>Class definition
    <th>re-usable instances
    <th>knot-tying instances
<tr><td><b class=mod>DefinedNames</b>
    <td><b class=mod>DefinedNamesBaseStruct</b>
    <td><b class=mod>DefinedNamesBase</b>
<tr><td><b class=mod>FreeNames</b>
    <td><b class=mod>FreeNamesBaseStruct</b>
    <td><b class=mod>FreeNamesBase</b>
<tr><td><b class=mod>NameMaps</b>
    <td><b class=mod>NameMapsBaseStruct</b>
    <td><b class=mod>NameMapsBase</b>
<tr><td><b class=mod>ReAssoc</b>
    <td><b class=mod>ReAssocBaseStruct</b>
    <td><b class=mod>ReAssocBase</b>
<tr><td><b class=mod>ScopeNames</b>
    <td><b class=mod>ScopeNamesBaseStruct</b>
    <td><b class=mod>ScopeNamesBase</b>
</table>

<p>
There are some additional modules in this directory:

<table><tr><th>Module<th>Purpose
<tr><td><b class=mod>PNT</b>
<td>A data type <b class=tcon>PNT</b>
    for identifiers carrying information on what unique entity
    it refers to, what type of entity it is (class, type, value, ...) and
    the position where it occurred in the source code.
<tr><td><b class=mod>UniqueNames</b>
<td>Data type <b class=tcon>Orig</b> (origin) to identity where something is
    defined.
    Data type <b class=tcon>PN</b> to tag an identifier with an origin.
<tr><td><b class=mod>ReAssocModule</b>
<td>A function for applying the reAssoc method to a module in the appropriate
    top level environment.
<tr><td><b class=mod>ScopeModule</b>
<td>Functions for applying the scoping pass to a module. The output is an
    abstract syntax tree containing identifiers of type <b class=tcon>PNT</b>.
</table>