- From: Adam Victor Reed <areed2@calstatela.edu>
- Date: Tue, 20 Nov 2001 17:01:25 -0800
- To: w3c-wai-gl@w3.org
Kynn Bartlett <kynn@idyllmtn.com> wrote: > What I'm thinking of, for example, are things like timed auctions -- > how do you make something accessible if they connect to an auction > site and there's 3 minutes left to bid on that item, and it will > take the user 5 minutes to comprehend the information and issue a > bid? > > According to both WCAG 2.0 draft (and 508, and Joe's proposal above), > there's no way to do it, really, and still be "accessible". In fact, > it may even be _true_ that there's no way to do it accessibly; it > may be a lost cause? > > I'm not saying I have the answer -- and in fact, my own attempts to > phrase things were self-admittedly poorly worded too. But I do think > that "timed events such as auctions" (or even "this sale lasts until > Tuesday" or the like) are what we need to look at as test scenarios > in addition to arbitrary ("keep your shopping cart!") or not-so- > arbitrary (e-learning testing) time limits. I think that there are 2 categories of time bounds which must be distinguished in our guidelines: 1. Universal, user-independent absolute deadlines (e.g. "This auction ends at 16:00 hours UTC.") For such limits there is no practical way, with currently available technology, to reconcile accessibility with fairness to other bidders. The guideline should specify that user-independent deadlines remain acceptable at this time. 2. Time-outs relative to user actions (e.g., your shopping cart vanishes if you don't submit complete billing information within five minutes after you received the check-out page.) The service will not qualify as "accessible" unless users can obtain an exemption from such limits. In situations (such as e-learning/e-testing) where response-time is used as a measure of performance, the fact that the user obtained such an exemption may be included in the subsequent report of the user's score. -- Adam Reed areed2@calstatela.edu Context matters. Seldom does *anything* have only one cause.
Received on Tuesday, 20 November 2001 20:01:33 UTC