Re: storing info in XSL-FO: new issue? [was: Draft TAG Finding:...]

Elliotte Rusty Harold wrote:
> At 10:28 AM +0200 8/20/02, Robin Berjon wrote:
>> The part I fail to understand is "even without agreement". How is that 
>> going to make more sense than a vocabulary that has agreed upon 
>> semantics such as MathML? Are we talking about a musician or about a 
>> sewing machine model?
> It will make more sense than MathML because we are not talking about 
> math, and MathML does not have the vocabulary to describe what we are 
> talking about (pop celebrities in this example).
> But the meaning is still in the message itself, 
> irrespective of whether I, personally, can usefully extract it.

Indeed, I agree that just because we can't decipher it doesn't mean that 
the meaning isn't there. The problem, however, is /accessing/ it. That's 
why we talk about *accessibility*.

There is no doubt that a specialised vocabulary can hold far more 
expressive power -- and thus potential meaning -- than a general one 
like XHTML.

Take the example of a long document contaning a mine of information 
about pop celebrities, and the user facing it only looking for a subset 
of that information (the bits about Madonna).

If that document is an HTML document in which sections, say about 
different pop celebrities, are marked with h1 headers containing the 
name of the said celebrity, a visual UA will allow the user to skim 
easily over the content to the place you're interested in. Similarly, 
that user could be using an aural UA that would be configured to read 
out all the h1's first so that he too could narrow down the useful 
(well, desired) information quickly.

First point: the same would not be true if instead of h1's that document 
had used solely visual markup. It would be impossible to differenciate 
between genuine section titles and user comment <p><font 
size='7'>Britney Sp34rs ROCKS, yay!!!</font></p>.

Second point: it might have worked had the document used PopStarML *and* 
if the user had access to a PopStarML UA. It might even have worked 
great with the UA humming a few notes of a popular song by each <singer> 
to skim over the list instead of reading the name. The problem is, 
functional (X)HTML UAs is already not a given, so new UAs for endless 
streams of new specialised vocabularies is unlikely to be workable.

However, unaccessible content is useless. We are talking about end-users 
here, not the ones that have fun reverse-engineering XML vocabularies on 
rainy sunday afternoons.

Your other post mentions (more or less directly) the possibility of UAs 
able to get at the semantics of a new vocabulary, and using it. That 
would be great as indeed if there were a way to associate 
ArbitraryVocabulary with Ontology with MediumNeutralRendering (in that 
order) then we'd have a much better Web.

Robin Berjon <>
Research Engineer, Expway

Received on Tuesday, 20 August 2002 10:00:08 UTC