forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Turner <je...@apache.org>
Subject Out-of-line links (Re: <acronym> tag in docv11)
Date Thu, 23 Jan 2003 12:11:34 GMT
On Thu, Jan 23, 2003 at 09:19:03AM +0100, Nicola Ken Barozzi wrote:
> 
> 
> Konstantin Piroumian wrote:
> >From: "Jeff Turner" <jefft@apache.org>
> >
> >>Howdy,
> >>
> >>Anyone mind if I add <acronym title="blah">blah</acronym> to
> >>document-v11.mod?  It's quite handy on occasion.
> >
> >
> >What's the use of it? Something like: <acronym title="ASF">Apache Software
> >Foundation</acronym> or the opposite <acronym title="Apache Software
> >Foundation">ASF</acronym>? How would it be displayed?

<acronym title="Apache Software Foundation">ASF</acronym>.  It's up to
the browser how to display it.  Mozilla (and I think IE?) have a light
underline and display the @title on mouse rollovers.

> Yes, I'm always confused by acronym, cite, blah...
> 
> Anyway, probably I'd like to see a glossary.xml that can keep terms, 
> acronyms, etc in a central place. The coolest thing would be not to have 
> to tag anything and have a transformer scan all the text for these 
> terms, but it's too time comsuming, so we still need a tag. IT could as 
> well be: <acronym>blah</acronym> or more generally <term>blah</term>
and 
> let the glossary.xml take care of making it into an acronym, word, or 
> whatever.

That would rock :)

It's not an either/or choice though.. for now, we can allow <acronym> in
the source.  Then later, we can implement a transformer that 'marks up'
raw text with tags like <acronym>.

The general idea of out-of-line linking, where you automatically mark up
raw XML, is pretty neat.  Some possibilities:

-  For jargon-heavy documentation, we could have a <term> tag for
   defining certain terms:

     "... cell proliferation and apoptosis..."
   ->
     "... cell proliferation and <term href="apoptosis.html" def="When
     cells go 'pop'">Apoptosis</term>..."

   That can be rendered as either "apoptosis?" (small superscript '?'
   link) or <span title="....">apoptosis</span>, ie a rollover
   definition.

   Again, we could start by defining an explicit <term> tag, and then
   later have a MarkupTransformer adding them for certain words.


 - We could implement Wiki linking by translating WordsLikeThis into "<link
   href="wordslikethis.html">words like this</link>".


 - Say we had a controlled vocabulary of geographical place names.  We
   could automatically mark them up with links to a map of the place:
 
     "... fires in Canberra, Australia ..."
   ->
     "... fires in <link
     href="http://www.maps.com/?lat=-35.3&long=149">Canberra,
     Australia</link> ..."

   that would be very useful for newswire services.
   
   <ot>
   At my old company (www.socialchange.net) we implemented this in
   an OpenGIS pilot project.  It requires some online GIS infrastructure
   though: geocoders, gazetteers, WMS mapservers, thesauri.. fun stuff..
   anyone up for a gis.apache.org? 
   </ot>


> >Maybe something like a <keyword>forrestbot</keyword> can be also useful
for
> >indexing and searching?

Docbook uses such things to build indexes.  If we could declare
taxonomies of words (eg 'pelican' is a typeof 'bird'), I'd imagine search
results could be more accurate.


--Jeff

Mime
View raw message