Re: Too classy? (Was: storing info in XSL-FO) wrote:
> However, it doesn't add what I termed "Structural Semantics". For 
> example, if a <singer> element were nested within a <CDCatalogue> 
> element or a <SewingMachineManufacturers> element we (but perhaps not 
> computers) get additional semantic information.

"CDCatalogue" as in "Couture et Design -- Catalogue"? Surely we're 
talking of a sewing machine here, no?

> Similarly, if the <singer> element had a namespace declaration 
> xmlns="" we get at least a hint of Namespace 
> semantics too.

Madonna is part of the Mafia?

> Do you believe that these additional dimensions of semantics can be 
> captured and expressed within a class attribute in HTML/XHTML?

You are talking about (english speaking) human beings. If that is all 
you want then meaning could be equally well defined in CSS comments 
relating to the used class.

But that's totally unrelated to the problem of preferring sufficiently 
well-defined vocabularies the elements of which have agreed-upon meaning 
over more ad hoc ones that provide no shared understanding. Eliotte's 
interesting but as yet unrealised idea of computers inferring sufficient 
meaning from visual rendering so as to be able to translate it to other 
media such as aural definitely makes for an interesting future. In the 
meantime, if we are to help accessibility we have to start by making 
things clearer for computers. That's why we still use things like 
keyboards and other such input devices instead of letting the computer 
read our brain activity and know what to do.

Robin Berjon <>
Research Engineer, Expway

Received on Tuesday, 20 August 2002 06:46:14 UTC