- From: Richard Schwerdtfeger <schwer@us.ibm.com>
- Date: Fri, 21 Sep 2007 14:22:50 -0500
- To: "Charles McCathieNevile" <chaals@opera.com>
- Cc: "Maciej Stachowiak" <mjs@apple.com>, "public-html@w3.org WG" <public-html@w3.org>, public-html-request@w3.org
- Message-ID: <OF186359DA.52A88F2E-ON8625735D.00696FA8-8625735D.006A7525@us.ibm.com>
Charles, I am on the same page as you that people will want roles for a lot of things and that is why we have namespaces in the XHTML ARIA specification. For html 5 we will be forced to remove the extensibility and it is a shame. Going forward we will have instances where people may create diagrams or who knows what. We allowed for different taxonomies to be build to be able discover and process them accessibly. Instead, we are forced to hardcode the set we have already defined. I agree the hint is great. We simply used alt. vs. label to do this. However, there is something very subtly here that command does not address that we do address with role (at least I don't see it). With role we would allow cycling among elements with common roles. I don't see how we can do that with label. So, here is a use case: Aria includes standard xhtml roles like main, secondary, and navigation. We could modify the command element to cycle through all the secondary landmarks (portlets) using the a key defined by the user agent or author. I don't see that with a label. Rich Schwerdtfeger Distinguished Engineer, SWG Accessibility Architect/Strategist Chair, IBM Accessibility Architecture Review Board blog: http://www.ibm.com/developerworks/blogs/page/schwer "Charles McCathieNevile" <chaals@opera.com To > Richard Sent by: Schwerdtfeger/Austin/IBM@IBMUS public-html-reque cc st@w3.org "Maciej Stachowiak" <mjs@apple.com>, "public-html@w3.org WG" 09/21/2007 02:02 <public-html@w3.org>, PM public-html-request@w3.org Subject Re: Proposal for Keyboard Shortcuts for HTML5 On Fri, 21 Sep 2007 20:24:19 +0200, Richard Schwerdtfeger <schwer@us.ibm.com> wrote: > Interesting. So there are few exceptions. I would like to see navigation > by role like access (when this makes it in). > > You could place an optional key assignment here like access and eliminate > accesskey altogether. Doing this would only work if you really have roles for everything. Given that the current move is away from having namespacing in the roles, based on the assertion that implementors will actually hardcode things rather than build an extensible mechnism, that might not be the obvious way to go about it. I agree that accesskey is inferior to most other ways of describing what something does and letting the system give it the default shortcut. But I am sure I can keep coming up with new things a widget will do faster than you can standardise and get implemented every kind of role - so having accesskey as a backstop that lets an author give a hint about what is meaningful would actually be valuable (plus it reminds more authors to think of keyboard access in the first place, with a bit of luck). cheers Chaals -- Charles McCathieNevile, Opera Software: Standards Group hablo español - je parle français - jeg lærer norsk chaals@opera.com http://snapshot.opera.com - Kestrel (9.5α1)
Attachments
- image/gif attachment: graycol.gif
- image/gif attachment: pic26517.gif
- image/gif attachment: ecblank.gif
Received on Friday, 21 September 2007 19:23:42 UTC