W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 1998

Re: introducing new semantics to HTML

From: Charles McCathieNevile <charlesn@srl.rmit.EDU.AU>
Date: Wed, 18 Nov 1998 17:23:18 +1100 (EST)
To: Jason White <jasonw@ariel.ucs.unimelb.EDU.AU>
cc: WAI Markup Guidelines <w3c-wai-gl@w3.org>
Message-ID: <Pine.SUN.3.91.981118171205.7300H-100000@sunrise>
MAP with block-level content is currently presented as if the MAP element 
were not there. The content is rendered, and the MAP is not related to 
the image. (I tested this with MSIE 4 and Netscape 4.07 and Lynx)

Which is a minor problem for users. On the other hand, both browsers 
tolerate the technically incorrect construction of a MAP which has both 
AREAs and block-level content. They render the content, and associate the 
AREAs with the MAP.

So I think it is OK to use MAP to cluster links, or to provide AREA-based 
imagemaps. I don't know if it violates the specification to have an 
imagemap which nothing refers to.

I would like to see MAP allowed to carry both blocks and AREA 
simultaneously. And I think I would like to see it as a block-level 
element. (I am not sure of this one.)

I would like to see UAs provide the content of A links in the same way as 
they provide the ALT-text of AREAs.

And then I would be a happy camper, suggesting that everyone's 
CSS2-compliant browser be  used with a stylesheet that does things which 
are appropriate to them with MAP and [USEMAP] and not worried about what 
classes they were.

Otherwise I think we should revisit the idea of CLASS="nav", which I 
think is an inferior solution.

cheers

Charles McCathieNevile

On Wed, 18 Nov 1998, Jason White wrote:

> All that is necessary is to indicate CLASS="nav" in the technique; then
> authors who implement the technique will use it, but it doesn't have any
> impact on HTML 4.0 as such. Quite simply, it becomes a widely used
> convention which default style sheets (or user-supplied style sheets) can
> follow, with confidence that authors who implement the technique will use
> it.
> 
> MAP is an interesting alternative, but we are yet to be presented with
> evidence of how existing user agents treat block-level content in this
> context.
> 
> On balance, I tend to prefer a conventional CLASS value, but I am
> obviously willing to be persuaded either way.
> 
> 
> 
Received on Wednesday, 18 November 1998 01:27:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:46:58 GMT