- From: Joshue O Connor <joshue.oconnor@cfit.ie>
- Date: Wed, 04 Jul 2007 14:44:36 +0100
- To: Robert Burns <rob@robburns.com>
- Cc: Maciej Stachowiak <mjs@apple.com>, Charles McCathieNevile <chaals@opera.com>, "public-html@w3.org" <public-html@w3.org>
A couple of brief observations/comments. > I think my main comment on accesskey would be that we should approach it like other features under recent discussion - start with the problem and requirements, not the solution Maciej has made this point a couple of times, and I have to say I agree. I find myself sometimes on this list (maybe due to the high volume of traffic) unsure of what point exactly, is being made/argued, where modifications are being suggest and in general where within the discussion issues are at. Again, it could just be a high traffic backlash. Anyway if I ever quote anyone out of context, please accept my apologies in advance. Rob said: > I don't think accesskey is meant to (nor should be required to) work on every device. >[...]So I think the accesskey merely needs to address the problems/use-cases it wants to address. I agree. What is accesskey for? Its usage on our desktop/laptop environments has been problematic due to *potential* for UA conflict, however this is often not as bad as it could be and real work accesskey disasters have by and large been few and far between. However, the fact that there was potential at all made them an undesirable mechanism for users to get around the page and I have always recommended that our clients remove them. Does the group thing there will be an about face in terms of their implementation in desktop environments? *If* accesskeys are to be used in order to facilitate easier user navigation of web interfaces then I also think, as Rob suggested, that they should be considered on a case by case basis with implementations modified according to platform, device user preferences etc. In terms of how they can be used on line (traditional desktop set up) they are pretty useless. Users of AT have other ways of navigating around and by groups such as this expanding the semantics available or using hooks etc future web interfaces may not need them at all. There can be situation where they are useful (as hacks to make up for a lack of hooks for example) and if they are to be used I suggest very much a constraint based approach, which limits the author defined key combinations they can use. Chaals, I haven't fully checked out your suggested specs yet but it looks interesting, will do so and comment forthwith. Josh
Received on Wednesday, 4 July 2007 13:45:11 UTC