- From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
- Date: Sun, 5 May 2002 21:42:18 +1000
- To: Web Content Guidelines <w3c-wai-gl@w3.org>
So fundamentally one could say it is a matter of using the most semantically rich and flexible technology available, all other considerations (i.e., user agent support) being equal. Specifically this would mean: Prefer style sheets to presentational markup Prefer text to images of text. Prefer vector graphics to raster-based images (again for reasons of scalability) Prefer formats that encode structure to those which do not. Allow the final presentation to be generated at that point in the delivery path at which software exists that can best satisfy the user's needs. Specifically this means in some cases that the presentational form can best be generated on the author's server, by a proxy, or by the user agent and that as content negotiation technologies improve the choice between these options will become more automated. Of course the countervailing consideration is, inevitably, user agent support, which will change in the course of time. Thus we can't freeze any particular "user agent baseline" into our guidelines, as it will soon be out of date, and in any event, developers shouldn't be penalized for using a newer, less supported but superior (from an accessibility point of view) technology by not being able to claim any level of conformance to accessibility guidelines. Cynthia's proposed solution might be of help here. Modifying it slightly, this proposal would be approximately as follows: 1. The content developer should define (or adopt from us) a user agent support baseline, e.g., HTML 4.01 + CSS level 1, with allowance for known implementation issues - or whatever it may be. This baseline should be documented by the developer and referred to from the conformance claim (especially a metadata conformance claim). 2. We would provide suggested baselines determined by whatever evidence we can gather regarding assistive technology support, and publish them periodically. 3. In the techniques documents we would have a consistent labeling scheme indicating, for each technique, the earliest version of the relevant technology in which the necessary features are available, with some comment as to known implementation status and any issues that developers should consider. Developers could then choose techniques according to whatever user agent baseline has been set. 4. Ideally, there should be tools for managing this process - that is, filtering techniques by technology version and/or implementation status. 5. The implementation status categories might be roughly as follows: a. not known to be implemented b. implemented, but not supported by assistive technologies and/or user agents typically used by people with disabilities c. Implemented by some user agents and, where applicable, assistive technologies, but not yet widely available. Internationalization/localization category: a. Not yet internationalized/localized b. Internationalized/localized by user agents c. Internationalized/localized by user agents, including some user agents/assistive technologies significant to people with disabilities. d. Internationalized/localized to a significant extent, and widely supported by implementations. Perhaps this is somewhat too complex and could be simplified a little - maybe a scale from 1-10 with points added or subtracted based on various factors (degree of user agent/assistive technology support, whether known implementation problems/incompatibilities exist, whether internationalized/localized versions are available), with fuller details included for those who want them. Disclaimer: I am not necessarily committed to any of the above ideas; they are just thoughts that might be useful in the resolution of the relevant issues.
Received on Sunday, 5 May 2002 07:42:33 UTC