README

              The HYpertext PARsing SUITE (Version 2)
              ---------------------------------------

 README
--------

This is a library meant mainly for information extraction purposes from *ML
documents. It takes a document as input, and creates a DOM tree from it, and
then allows you to work with it. It's very tolerant towards malformed documents
(that is, documents which do not well-formed according to the XML terminology.)

The library needs a table containing rules which tell it which tags are valid,
and where those tags should occur. It's completely table-driven, without any
hard-coding of rules. There is a default table containing rules which resemble
the HTML-rulesets very closely (the table has its limitations - it can not
capture the rules which DTDs can specify perfectly ...)
If no table is supplied, default 'XML well-formed'-ness rules are applied.

For the sample rules-table of HTML, look for the file htmltable.cpp within the
src/hypar directory.

The test/ directory contains a lot of sample code for testing various
components and serves as a good starting point to figure how to use the
library.

-- 
All copyrights reserved by Ravindra Jaju (2004 -)
<lastname-in-smallcase> [AT] msync org