Re: introducing new semantics to HTML

Hmm. So when are you going to send us some of these poems? *smile*

To the point:

CLASS and REL are similar beasties. Both allow extension of semantics. 
REL has some values which are suggested by W3C (eg stylesheet) and 
probably lots which are introduced by people building their intranets out 
of HTML and fancy parsers. But where a REL value is widely accepted, this 
is a Good Thing (TM), as it allows for reproduceable functions to be 
added to HTML.

If there were some (more or less) standard CLASS values then we could 
reuse stylesheets as well. But as Nir says, if we want to reserve values 
we need to give a long lead time. (Much as there is a long lead time 
before the reserved terms LONGDESC, OBJECT, etc are actually really useful.)

In the case we are discussing, we are looking for a way to deal with 
clusters of links. the MAP element was designed to do so. In the 
transition from HTML 3.2 to HTML 4.0 it was made much better at it. In 
the next transition there may be a few more improvements (allow AREA and 
block-level content, make it block level?). At that point the WAI may 
wish to suggest that several CLASS values be 'reserved'.

To allow browsers to set them as display:none would be more than I was 
happy with - I would prefer for defaults to simply shift them to the end 
of a page. But given the amount of lead time before any of this is likely to 
actually be implemented, the whole prospect does not concern me unduly. 
It is important that IF we look at this path then there is an onus on EO 
to ensure that this is well known, Internationally.

An alternative would be to allow an attribute NAV to the element MAP. But 
it does seem redundant - MAPs made of AREAs don't have any initial 
rendering (although good things like lynx offer the possibility) and MAPs 
are collections of links - I find it hard to think of an example where 
this is not a navigation system. Not that that means it doesn't happen.

So although in principle I think that we can reserve a CLASS value or 
two, in practice I do not see that it is necessary.

Charles McCathieNevile

Received on Tuesday, 17 November 1998 20:38:08 UTC