$Id: README,v 1.6 1997/10/11 00:01:48 ken Exp $


                            SGML::SPGrove

                   A Perl 5 module for loading SGML,
                   XML, and HTML document instances
                       using James Clark's SP.


                             Ken MacLeod
                         ken@bitsko.slc.ut.us

INTRODUCTION

    The SGML::SPGrove module links with James Clark's SGML Parser (SP)
    to load SGML, XML, and HTML document instances.  SPGrove supports
    the Visitor design pattern for accessing the grove and also comes
    with a module for performing simple rule-based transformations to
    Perl objects.

    See the file Changes for user-visible changes.  See the `examples'
    directory for examples of using `SPGrove'.

    Newer versions of this module can be found at
    <ftp://ftp.uu.net/vendor/bitsko/gdo/>.

    Copyright (C) 1997 Ken MacLeod
    SPGrove is distributed under the same terms as SP.  See the file
    COPYING for distribution terms.

OVERVIEW

    SGML::SPGrove takes a system identifier and passes it to SP to
    parse, as each element is parsed from the document SPGrove builds
    Perl objects to match.  When done parsing, SPGrove returns an
    SGML::SPGrove object that contains the root element of the parsed
    document and an array (hopefully empty :-) of parser errors.

    Elements of the document are SGML::Element objects.  Elements
    have a generic identifier (or name), attributes, and the contents
    of the element.  Attributes are stored as a Perl hash, with the
    values as an array of scalars and SGML::SData objects.  The
    contents of an element may be more Elements, scalars, SData
    objects, or processing instruction (PI) objects.

    SGML::SData objects are replacements for character entity
    references within the document.  The Text::EntityMap perl module
    can be used to map SData replacements from common character entity
    sets to common output formats.

<ftp://ftp.uu.net/vendor/bitsko/entity-map/>

    SGML::PI objects are processing instructions contained within the
    document.

    SGML::Simple::Spec, SGML::Simple::SpecBuilder, and
    SGML::Simple::BuilderBuilder work together to implement a simple
    rule-based transformer for transforming document instances to Perl
    objects.  SpecBuilder takes a spec grove conforming to the
    ``SPGrove Simple Spec'' DTD and creates a specification object
    that can be given to BuilderBuilder to create a Visitor package
    that can be used to transform other groves to Perl objects.

    See the pod/man pages for SGML::SPGrove, SGML::Element,
    SGML::SData, SGML::PI, and SGML::Simple::Spec for more info.

    Visitors and Builders are explained thoroughly in ``Design
    Patterns: Elements of Reusable Object-Oriented Software'' by
    Gamma, Helm, Johnson, and Vlissides, published by Addison-Wesley
    (ISBN 0-201-63361-2).

<http://st-www.cs.uiuc.edu/users/patterns/patterns.html>

INSTALLATION

    SGML::SPGrove requires Perl 5 and James Clark's SP (from the Jade
    distribution).  SP requires a C++ compiler.

<http://www.perl.com>
<http://www.jclark.com>

    1)  SPGrove needs SP's `libsp.a' and include files.  SP's `make
        install' does not install these [I'm working on that, I should
        have an RPM available soon].  Create a workarea for compiling
        SP, compile it and keep the workarea until SPGrove is done.

        Edit SPGrove's Makefile.PL to point `LIBS' to SP's `lib'
        directory and `INC' to SP's three include directories.  I've
        left my templates in to point the way.

    2)  standard Perl module after that,

            perl Makefile.PL
            make
            make test
            make install

        Just so you know, SPGrove's copy of SP's library is included
        in the install, that's over a megabyte and a half.

        FYI, a statically linked perl executable (`make perl') appears
        to run significantly faster, in one test, 17 seconds instead
        of 25 seconds.