- From: Matt May <mcmay@bestkungfu.com>
- Date: Mon, 25 Dec 2000 15:19:54 -0800
- To: <w3c-wai-gl@w3.org>
----- Original Message ----- From: "Sean B. Palmer" <sean@mysterylights.com> > > I can't see a problem with requiring one standalone RDF > > document for compliance checking, title/author/description > > While I fully support this notion as a technique (obviously, else I would > be working on UWIMP etc.), I think that it would be hard/inadviasable to > outline implementation details in the WCAG guidelines (especailly as one of > the key tenets of WCAG is technology independance). It would be best to say > something about adding indexing according to technological contraints, and > then adding a fuller eplanation to an XHTML/RDF Techniquies document. If the guideline were to say, "use indexing technologies to (foo)", I think that, at least, the techniques document would have to say "by indexing technologies, we mean RDF, done like this," until something better comes along. The technology independence plank, I believe, is to allow for entropy in data. Entropy in metadata is a whole other ball of wax, potentially preventing the proposed guideline from achieving its desired effect, and if we don't say what we mean here, that's what we'll get. > > I may have issues with embedded meta tags in HTML > > documents, however. > > Could you possibly iterate these issues (if you believe they'll be of > benefit to the list)? I can think of two offhand: - Labor-intensive for users with legacy content. I'd prefer that user agents that need or desire WCAG compliance metadata look for it in a certain place (e.g., ./wcag.xml), and designers can override it if they like with a <link>. (A perhaps welcome side-effect here is that non-compliant content providers will begin to see how many users they have with assistive technologies, simply by tracking 404s.) - meta name="description" is already used (read: contaminated) by search engines. The contents of the average meta description tag are questionable because many content providers exploit them for best results in the engines, not for accurately describing the data within.
Received on Monday, 25 December 2000 18:19:57 UTC