- From: Charles McCathieNevile <charles@w3.org>
- Date: Sun, 20 Aug 2000 07:35:13 -0400 (EDT)
- To: "m. may" <mcmay@bestkungfu.com>
- cc: Jason White <jasonw@ariel.ucs.unimelb.edu.au>, Ian Jacobs <ij@w3.org>, Web Content Accessibility Guidelines <w3c-wai-gl@w3.org>
Here is a very slight modification to that proposal: We provide a list of checkpoints, which are the abstract requirements for accessibility - for example that there is a text equivalent of all information. It is important that they are testable. But they don't need to be machine testable since that may not be possible. These abstract checkpoints are part of a Recommendation. We should also provide technology-specific ways of meeting the checkpoints. Since these are going to change more often, and since they are unlikely to be unique in most cases (there are several ways to provide text equivalents for an image in HTML 4.0. On the other hand, there is more or less one way to ensure that code is syntactically valid) these might as well be a note. Each of these methods (which I propose we call "Techniques") may be a way to completely satisfy one or several checkpoints, or it may be a way to partially satisfy a checkpoint. This information should be part of the technique. That way, conformance is always to the checkpoints of the Recommendation, but there is a clear way to achieve it for a given technology - by choosing techniques that meet all the checkpoints. This proposal has the added advantage of nobody having to learn a new way of dealing with WAI guidelines, since that is how it works at the moment. And even the terminology remains stable. cheers Charles McCN On Sat, 19 Aug 2000, m. may wrote: On Sat, 19 Aug 2000, Jason White wrote: > I like Ian's proposal, a variation of which might well work. One problem, > however, is as follows: > > Suppose that an individual or organisation either develops a new > technology, for example a markup language, or extends an existing > technology (noting that most of the recent W3C formats are designed to be > extensible). Further, let it be assumed that the W3C has not developed a > profile for this technology (or, in the case of an extension, the profile > does not take account of the added features). Two questions arise: Let me add a twist here. Let's say that a new, uncovered, but theoretically and practically accessible technology is introduced. Through pressure, political or otherwise, organizations adopt WCAG 2.0 as an accessibility baseline, and require all browsers/apps, tools and deliverables accommodate the guidelines. In the absence of valid technology-specific checkpoints, these organizations aren't allowed to use the technology. This puts the W3C in the precarious position of keeping tools out of the hands of potential customers. That stands a chance of becoming a political hot potato, and the W3C may risk losing the support of some affected vendors. My proposal: - A list of general checkpoints to satisfy the requirements of accessibility which is appropriately abstract and extensible. This is a W3C Recommendation. - Technology-specific checkpoints, as appropriate, delineate how to completely satisfy goals of accessibility in each technology. This is a W3C Note. (My understanding is that Notes are easier to revise and extend; I feel that's necessary for this kind of document. If I have it wrong, that extensibility is my intent, anyway.) - Compliance is based on the tech-specific checkpoints where they exist, and on the general checkpoints where they do not. I think we also need to remember that many of the techniques being discussed are, essentially, errata: they're fixes for holes that have been exploited in the technology's specification. As such, the tech-specific documents need to be living documents, and thus shouldn't be subject to the W3C recommendation process. And in that vein, another issue I'd like to raise comes from something Jason mentioned: Many recent W3C protocols are designed to be expanded, and the resulting products may not map directly to an already-published technical doc. For example, XML can be _anything_. Accessibility when it's used as RDF means a whole different thing from when it's used as XHTML, or simply as an object database. I think there's going to be an awful lot of hair-splitting when it comes down to implementation because accessibility rules change not just by technology, but just how it's used. In the absence of a spec for everything under the sun, requiring conformance with the higher-level rules seems like a fair compromise. Anyway, that's my spin. :) ---- matt -- -- Charles McCathieNevile mailto:charles@w3.org phone: +61 (0) 409 134 136 W3C Web Accessibility Initiative http://www.w3.org/WAI Location: I-cubed, 110 Victoria Street, Carlton VIC 3053 Postal: GPO Box 2476V, Melbourne 3001, Australia
Received on Sunday, 20 August 2000 07:35:22 UTC