From lhs at malform.no Sun Sep 21 17:25:05 2008 From: lhs at malform.no (Leif Halvard Silli) Date: Mon, 22 Sep 2008 02:25:05 +0200 Subject: [html4all] Longdesc and data URIs Message-ID: <48D6E5E1.70502@malform.no> Gallery2 is a GNU licensed PHP gallery software which (faulty) longdesc support. http://en.wikipedia.org/wiki/Gallery2 Demo page with longdesc: http://gallery.menalto.com/gallery/gallery-team/ However, as is evident from that same page, Gallery2 has gotten it wrong ... Instead of a URL, its treats the attribute as as a text attribute: It is interesting to consider how Gallery2 could have improved their software: * Data URI: The text inside the longdesc attribute could be converted to a data URI - should be an simple conversion. Of course, data URIs isn't necessarily compatible with current AT software. However, we must expect that AT software will read data URIs one day. * same page longdesc: The longdesc texts could be moved to a hidden "footnote area" on the same page. * separate page longdesc: And of course, they could link to longdescs on a separate page - in the tradtional way. I tested the data URI solution, and it works fine (via JavaScript) in Webkit, Opera and (I suppose) Firefox. But not in Internet Explorer 8 beta 2 - as Microsoft claims it is a security risk to support data URIs for anything but binary data. -- leif halvard silli From laura.lee.carlson at gmail.com Tue Sep 23 06:21:14 2008 From: laura.lee.carlson at gmail.com (Laura Carlson) Date: Tue, 23 Sep 2008 08:21:14 -0500 Subject: [html4all] Article: Web Standards 2008: Three Circles of Hell Message-ID: <1c8dbcaa0809230621j32877805vaf2ac1cc1d75d82c@mail.gmail.com> Web Standards 2008: Three Circles of Hell By Molly E. Holzschlag. "Standards promised to keep the web from fragmenting. But as the web standards movement advances in several directions at once, and as communication between those seeking to advance the web grows fractious, are our standards losing their relevance, and their ability to foster an accessible, interoperable web for all?" http://www.alistapart.com/articles/webstandards2008 Best Regards, Laura -- Laura L. Carlson From foliot at wats.ca Tue Sep 23 10:58:15 2008 From: foliot at wats.ca (John Foliot - WATS.ca) Date: Tue, 23 Sep 2008 10:58:15 -0700 Subject: [html4all] ARIA (Was RE: function and impacts (was: @scope and @headers reform)) Message-ID: <00d301c91da5$f63493e0$e29dbba0$@ca> Laura Carlson wrote: > > > Gez wrote, > > > If HTML5 allows > > aria-labelledby on table cells, I would be in favour of dropping > the > > headers attribute completely, as it can't be used practically, and > use > > aria-labelledby for complex associations instead. > > This is an interesting proposal and perhaps a way forward. > > The question is will aria-labelledby on table cells be written into > the html5 spec? All, Recently, I had an opportunity to talk with TV Ramen, who you may have noticed has remained pretty much outside to all HTML 5 discussions. We talked at length about @alt, @longdesc, and the myriad of other gaffs and omissions we are currently enduring. His advice was to focus instead on ensuring that ARIA and its implementation(s) are robust, complete and fully featured, as despite the flavor of HTML being authored, if @aria attributes are used, the AT user will benefit. (For example, ensuring that appropriate @role values exist for images) After all, we already have browser implementers supporting and extending ARIA with each coming day. In a way, it's kind of sad to ghetto-ize specific accessibility functionality into a sub-set of HTML, but given that this appears to be where the real gains are being made, perhaps we need to take the pragmatic approach of advocating for the inclusion of all of ARIA's work into HTML5 wholesale (now, or down the road when things have settled down) and instead focus on ensuring that all of the required functionality we see and know is needed is addressed through an ARIA method - in other words find real solutions outside of the cabal's machinations and seek to import "en-bloc" the entire ARIA toolkit as part of the final official spec. Thoughts? JF From lhs at malform.no Tue Sep 23 12:44:17 2008 From: lhs at malform.no (Leif Halvard Silli) Date: Tue, 23 Sep 2008 21:44:17 +0200 Subject: [html4all] Do we need axis or headers='a-TD-cell' ? Message-ID: <48D94711.5060409@malform.no> In HTML 4, in the context of the HTML 4 heading algorithm, axis="" and headers="a-TD-cell" are equal sizes: Both let a certain cell (either the one with the axis="" attribute or the one which headers="" points to) be incorporated into the @headers incorporpoating heading algorithm of HTML 4. Now, in HTML 5, if the purpose of @axis and headers="a-TD-cell" is *only* to label certain TD cells as header cells, *without* making them look like bold styled TH cells, then the purpose of both axis="" and headers="a-TD-cell" is purely stylistic. One would then be able to replace the following (Ex. 1):
cell.
[1] http://lists.w3.org/Archives/Public/public-html/2008Sep/0429.html
--
leif halvard silli
From P.Taylor at Rhul.Ac.Uk Thu Sep 25 05:09:59 2008
From: P.Taylor at Rhul.Ac.Uk (Philip TAYLOR (Ret'd))
Date: Thu, 25 Sep 2008 13:09:59 +0100
Subject: [html4all] Grandfathering ?
Message-ID: <48DB7F97.3070401@Rhul.Ac.Uk>
I have little or no understanding of the underlying
issues concerning table headers and IDs, so I have
steered well clear of this discussion, silently
siding with any and all suggestions that can
improve the accessibility of tables. However,
in trying to catch up with the thread, I encountered
the following from Henri :
> In the part of my email you didn't quote, I said I was in favor of grandfathering headers/id into the spec.
Could anyone (I imagine it would have to be a native
|