- From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
- Date: Wed, 4 Sep 2002 19:34:13 +1000
- To: Web Content Guidelines <w3c-wai-gl@w3.org>
Ben has raised a number of significant questions in his summary. By way of contribution to this discussion, I have several personal opinions to offer, as usual. These are placed before the group for discussion and debate, as necessary. 1. I think "dynamically-generated" presentations would be preferable, if they can be implemented given the resources available to the group. For example, if I am evaluating a web site that uses a given combination of technologies, I should be able to generate a checklist that specifies all of the techniques relevant to my situation, with references to the success criteria that they satisfy. Alternatively, if I am writing an authoring tool which implements a given technology and I want to ascertain what techniques should be followed in order that my tool can produce highly accessible output that meets WCAG 2.0, then I will need a checklist that is restricted just to the technology in question. 2. Ben inquired how technologies which, individually, aren't relevant to all of the checkpoints should be treated. One solution would be, instead of taking each technology separately, to write techniques which presuppose that certain technologies are used in combination with each other. For example, XHTML is typically used with one or more audio or image formats, and with a style language (usually, but not always, CSS), so it might be better to combine the relevant techniques and examples under the assumption that certain types of technology are typically used together. If this combining could take place automatically it could be very flexible indeed, and such a possibility might be worth investigating further. CSS is a good example of a technology for which there are techniques, but it is always used in conjunction with XML or HTML. The conformance claim would be made in respect of the combined XML/XHTML/HTML/CSS etc., content; and it should be easy for an author to obtain a checklist that recognizes this fact. I don't want to work through one checklist for XHTML, another for CSS, etc., but rather to complete a single, unified checklist relevant just to the technologies which I have chosen. Whether this desire can be practically managed remains an open question, but we should at least consider the issue. 3. Regarding the review requirements in the level 2 success criteria, by definition there aren't any "hard and fast" rules for determining whether the content has the indicated characteristics; this is why the success criteria say that a review must be carried out (taking into account the information we provide), and, on that basis, an opinion formed. The review should be undertaken according to a consistent process and may involve testing with actual users (cf. Appendix B of the latest internal draft). We may be able to provide advice in the techniques on how to conduct such reviews. Beyond that, I don't think definite technology-specific techniques can be laid down; if they can, then they ought to serve as the basis of further success criteria in the guidelines proper. The Education and Outreach working group, as I recall, has been working on issues of content evaluation and may have advice to give regarding "best practice" in the conduct of qualitative reviews. Also, we can offer (technology-specific) examples of what we regard as exemplary. 4. As already indicated, I think we should distinguish carefully between techniques and examples. When Wendy and I reviewed the guidelines last November to gain an overview of how they related to the techniques, we discovered that there were many issues best addressed in "core" techniques, without referring to any specific technology, but for which we could also provide technology-specific examples. Thus, an explanation of how to write a good text equivalent (checkpoint 1.1) need not refer to any particular technology, but we can certainly provide examples of how checkpoint 1.1 as a whole is implemented in specific technologies, for example SVG or XHTML. There aren't so many points in the guidelines at which technology-specific advice is needed (what I mean is that they tend to relate to certain checkpoints in the docment, e.g., under guidelines 1 and 2, less so in guidelines 3 and 4, and again in guideline 5, to make a broad generalization here). 5. It has sometimes been suggested that the WCAG 2.0 techniques should state explicitly which techniques, in which combinations, are regarded as satisfying specific success criteria, and which ought instead to be considered as alternatives. To be specific, implementors want to know, for example, that a given success criterion can be satisfied by implementing either technique A, or techniques B and C together. This kind of information is essential. I don't have a firm opinion on whether alternatives should be prioritized. Obviously, selection among alternative techniques is likely to be governed not only by technological factors (issues of user agent baseline, checkpoint 5.2) but also the author's design preferences more generally. I would argue that each technique which relies on a particular feature of a technology should indicate, at least, the version of the technology in which the relevant features first became available. Whether filtering can be carried out with respect to this information depends on how we set up the dynamic checklist generation system, which appears to me increasingly to take the form of a data base with a prescribed set of fields for each technique. I think that covers many of the issues which Ben raised.
Received on Wednesday, 4 September 2002 05:53:33 UTC