- From: Sander Tekelenburg <st@isoc.nl>
- Date: Sun, 23 Sep 2007 20:45:45 +0200
- To: <public-html@w3.org>
At 14:33 +0200 UTC, on 2007-09-22, Charles McCathieNevile wrote: > On Sat, 22 Sep 2007 03:05:05 +0200, Maciej Stachowiak <mjs@apple.com> > wrote: [...] > Given that nearly all browsers provide no way at all to discover that > there are accesskeys (even when they implement them in a way that lets > them hijack the standard UI) I think the 2% figure is surprisingly high. Indeed. Note also that some Automated Web Authoring Tools generate @accesskeys. No doubt that notably ups such numbers. - The default template of CMSMadeSimple contains accesskeys for "skip to content" and "skip to navigation", for example. - The default template of Dokuwiki contains lots of accesskeys for common navigative elements like go to home page; create page; show old revisions; show recent changes; search; go to index; etc. - Both FCKEditor and TinyMCE let authors enter accesskey values when creating links. (However, TinyMCE happily swallows multiple characters. At least in the version I tried.) [...] >> Since the legal values, expected behavior, and set of elements where it >> would be allowed are all different, I don't think reusing the name would >> provide compatibility with the web. > > For the case where you come up with only one value it seems to provide > compatibility. For the case where you decide you want two or more, I agree > - but then it is currently undefined what happens in that case, which > seems to apply to a bit less than 0.02% of the web, and it would be > helpful to specify the error handling. Note that iCab doesn't choke on multiple characters in @accesskey. It accepts the first character and ignores the rest. So allowing a list of accesskeys would be backwards compatible with at least one UA. (I haven't tested others.) [...] -- Sander Tekelenburg The Web Repair Initiative: <http://webrepair.org/>
Received on Sunday, 23 September 2007 18:54:02 UTC