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

17.08.02 15:03:06, Elliotte Rusty Harold <elharo@metalab.unc.edu> wrote:

>No, I'm saying that encouraging better use of HTML is a waste of 
>time. The battle's lost. XHTML 1.0 has achieved no significant 
>traction. XHTML 2.0 will probably do worse. If you want semantically 
>tagged information, jump to pure XML; but accept that HTML is today 
>and probably always will be a layout language.

XML has *no* semantic information, (X)HTML has *some*. If you know a 
resource is XML, you know a little about the resource. You know that it is a 
tree structure, you know that two elements with identical name may share 
some sameness, and so on. You don't know what those elements, attributes and 
characters mean. Validation adds important syntactical assurances, you may 
know that a <h1> may not contain another, but you still won't know what it 
means. In (X)HTML, you may know it is a top-level heading and you may expect 
that this was what the author intended it to be.

XSL-FO doesn't subtract from nothing, but it doesn't add much either. 

SVG is a very rich graphics format, but it is a graphics format and 
inappropriate as a document format. It would be like using Adobe Photoshop 
as a text editor, and marginally better than having a web site consisting of 
screen shots.


>What the W3C says about HTML has little effect on what web sites and 
>designers do. The language of HTML is defined by the tools, not by 

Agreed, HTML is messy because it is in use, so would any other W3C 
specification be. Like with a building, a specification is only untainted 
and clean when it is an architect's drawing. HTML has been lived in for a 
decade, at times it has been squatted, but is a thriving and mostly 
functional building. If it had been remodeled into a cathedral, where the 
tenants are forced to listen to a weekly sermon, it would have been deserted 
in no time.

The W3C specifications matter, otherwise none of us would be here. Since 
there are millions of people using HTML, most who never have been to the W3C 
site and never will, there will always be some "genetic drift". People may 
be using features for purposes they were not intended for, not using 
features they were intended for, or use features elsewhere than intended. 
But people who make the tools, or the books and articles, do consult W3C. As 
long as W3C makes good specifications, they will be listened to. Some 
requirements would be

(*) The features must be easy to understand
(*) They must be easy to use right (and preferably hard not to wrongly)
(*) It should be easy to know right from wrong
(*) The features must fulfil a need


Jonny Axelsson,
Documentation,
Opera Software

Received on Sunday, 18 August 2002 08:08:25 UTC