Re: What are Semantics? (Was: Serving generic XML)

On Monday 2002-08-19 15:56 -0400, Svgdeveloper@aol.com wrote:
> So, according to Kynn, generic XML has no semantics.
> 
> Interestingly, he views a stylesheet as supplying semantics.
> 
> Yet he claims that XHTML possesses "semantics".
> 
> But what are the characteristics of such "semantics"? Are these 
> presentational pseudo-semantics only?
> 
> Other views appear to assume that there is "semantic markup" - presumably XML 
> - which is "rich" in "semantics" and presentational markup. Interpreted in 
> some ways a view that is diametrically opposed to Kynn's since he suggests 
> that generic XML has "no" semantics.
> 
> Just as I was about to post this email, a post from Elliotte provided a scale 
> on which XML had the "most" in terms of semantics.
> 
> It seems to me that we need to be clearer about terminology since two, 
> seemingly intelligent, individuals interpret the semantic richness and 
> poverty of XML in two diametrically opposed ways.

The semantics that are relevant for web accessibility are those that
allow markup to be presented appropriately in different media, to
humans.  For example, a bold font (or a large font) for the screen
medium might represent a heading, which might be represented in an aural
medium using pauses and speed changes, or it might represent emphasis
that is best represented in an aural medium using changes in volume or
pitch.  Likewise, indentation in a screen medium might represent a
tree-like structure of a nested list, which might be best represented in
an aural medium using additional words to describe the structure, or it
might just be stylistic indentation, best ignored in an aural medium.

These are not the only semantics that are relevant for the web.  For
example, slightly different semantics are useful to search engines,
which would find it useful to know that something is a heading (and thus
describes a large section of the document) but probably don't find the
distinction between lists and paragraphs as relevant.  However,
considering the wonders of Google, I'm more worried about web
accessibility these days than searchability.

"Generic XML" is not useful for either of these things, since the
reciever doesn't know the vocabulary.  (Even if it is English, computers
aren't very good at English.  Hopefully you don't expect the average web
*user* to be reading your *markup* directly, do you?)  XHTML is useful
for both.

-David

-- 
L. David Baron        <URL: http://www.people.fas.harvard.edu/~dbaron/ >

Received on Monday, 19 August 2002 16:33:03 UTC