Aaron Straup Cope <asc@[...].net>  writes:

>  http://search.cpan.org/author/ASCOPE/XML-Filter-Glossary-0.1
>  
>  "This package is modelled after the UserLand glossary system where
>  words, or phrases, wrapped in double-quotes are compared against a
>  lookup table and are replaced by their corresponding entries.

>  "Keywords are flagged as being any word, or words, between double
>  quotes which are then looked up in the glossary. If no match is
>  found, the text is left unaltered.

The UserLand glossary system is well known for catching
false-positives -- matching words the user didn't know were in the
glossary.

One often suggested alternative is to use a glossary-namespaced XML
element in place of the plain text and then use empty (or not??)
glossary elements to call out glossary items:

  $glossary-> set_glossary("/usr/home/asc/bookmarks.xbel");
  $parser-> parse_string("<?xml version = '1.0' ?><root xmlns:g="http://aaronland.net/glossar
y"> This is <g:aaronland/></root>");

  # prints :

  <?xml version = "1.0" ?> 
  <root> 
   This is <a href='http://www.aaronland.net'> aaronland</a>
  </root> 

In the start_element() of XML::Filter::Glossary, you can then do:

  if ($element-> {NamespaceURI} == 'http://aaronland.net/glossary') {
    # merge content based on $element-> {LocalName} as the key
  }

Probably much simpler and far more foolproof.

For user-entry boxes, like posting and comment textinputs, you can
"presupply" the xmlns:g attribute in the wrapper for their entered
text (if xml) or do a literal search for <g:WORD/>  and handle
appropriately.  For users who are aware of the glossary, <g:WORD/>  is
no difference, for users unaware they can never get hit by the
glossary by accident.

  -- Ken
_______________________________________________
Perl-XML mailing list
Perl-XML@[...].com
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs