- From: Leif Halvard Silli <lhs@malform.no>
- Date: Tue, 19 Aug 2008 20:02:52 +0200
- To: "Patrick H. Lauke" <redux@splintered.co.uk>
- CC: Sam Kuper <sam.kuper@uclmail.net>, "public-html@w3.org WG" <public-html@w3.org>, W3C WAI-XTECH <wai-xtech@w3.org>
Patrick H. Lauke 2008-08-19 00.52: > Sam Kuper wrote: > CMSs and authoring tools, in view, are a completely separate beast. In > my last email I noted how many users (myself included) don't user Flickr > so much as a publishing platform, but as an online extension of their > private photo collection, with the ability to share with known > individuals, friends, family. The use context is different. In the case > of CMSs used for publication to the wide web, I agree - and many more > sophisticated systems already do warn about empty alts - but users > should still be able to consciously leave alt empty, as otherwise > they'll simply enter some generic text ("photo", "image", "abc", "blah", > ...) if they're forced to enter metadata/alternatives. I like these viewpoints, and the example in another letter about whether a table without header cells etc should validate at all, was well taken. But how can we differentiate between public and private? My thought is that the role="" attribute could play a -eh- role here. One could have special requirements for the @alt, depending on its role="", basically. If e.g. role="an-private-album-photo" had more relaxed rules, for instance that it said that the @alt /could/ be emtpy then, then it would be possible to fake validity by using that @role value e.g. in a government web site. (And the same would be possible by always using role="decorative".) However, it would then be simple to spot the fakery. And also, one would have better chances at developing a sound understanding of what "validity" means. If we are to leave the whole issue of the @alt requirements to the WCAG specification(s), then all we need to do is to take the @role into HTML 5, and ask that authors use the @alt according to the requirements which the WCAG set forth for each role. Even defining the values of the @role attribute, could then be "outsourced". (Allthough I think that something should be said about how User Agents should handle e.g. <img alt="spacerimage" role="decorative" src="src" >.) Btw, note that the current values of the @role attribute appear to be quite "dry". It seems to me that they are very "public", with very very little focus on e.g. the Flickr use case. -- leif halvard silli
Received on Tuesday, 19 August 2008 18:03:41 UTC