RE: Accesskey: there are "techniques"?

To second what Tom is saying - in my view the problem with access keys
conflicting with browser accelerator keys is a user agent problem. The
problem is that the same modifer key is used to activate keys that are in
separate "namespaces", the browser namespace and the document namespace. The
solution is to use a different accelerator key for the access keys defined
in HTML, then the entire set of keys is available. I believe Opera does
handle access keys in this way and we're just waiting for other browsers to
catch up. The issue of discoverability remains but is a separate issue.
While we need to log user agent issues with our techniques, I think we
should avoid coming up with recommendations that will hopefully become
invalid at some time in the near future. Michael

> -----Original Message-----
> From: Tom Croucher [mailto:tcroucher@netalleynetworks.com]
> Sent: Friday, September 26, 2003 8:59 AM
> To: 'Yvette P. Hoitink'; 'WCAG List'
> Subject: RE: Accesskey: there are "techniques"?
> 
> 
> 
> Surely this is user agents issue rather than an accesskey issue. Opera
> uses an accesskey mode which turns off its default shortcuts while in
> that mode. Since you can toggle the mode on and off you know exactly
> what you are getting. It seems to me that developers shouldn't feel
> hamstrung by any implementation (however large the IE market slice is)
> of accesskeys in UAs, there are 36 keys to use and we should use them.
> On complex sites having an intuitive letter for the 
> accesskeys can make
> a big difference.
> 
> My two pence.
> 
> 
> Tom
> 
> Co-founder Netalley Networks
> (http://www.netalleynetworks.com),
> BSc(Hons) Computing Student / Information Services Staff University of
> Sunderland
> (http://www.sunderland.ac.uk),
> Accessibility Co-ordinator Plone CMS
> (http://www.plone.org)
> 
> -----Original Message-----
> From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On
> Behalf Of Yvette P. Hoitink
> Sent: 26 September 2003 13:48
> To: 'WCAG List'
> Subject: RE: Accesskey: there are "techniques"?
> 
> 
> Roberto Scano asked:
> > There are some techniques about a "map" for the accesskey?
> > 
> > eg:
> > 
> > 0 - Home Page
> > 1 - Access Key page
> > 2 - Site Map
> > 3 - Contact e-Mail ...
> 
> One big problem I always have with accesskeys is that there 
> is a risk of
> 'hijacking' key combinations that are already in use by the user. For
> example, Alt-H opens the Help menu in an English version of Internet
> Explorer. But when I attach H as an accesskey, suddenly Alt-H can no
> longer
> be used to open the Help menu. Also, especially people with 
> disabilities
> may
> have assigned keyboard shortcuts to functions they use a lot 
> which would
> get
> overridden. Adapting another persons environment is _not_ what I call
> accessibility. 
> 
> For this reason, I only use numbers as accesskeys. I tend to use 1 for
> the
> start page, 2 (to) for the skiplink (which goes directly TO 
> the content)
> and
> 4 (for) for the searchbox. I use 5-0 for the main navigation links .
> This
> leaves me 3 another important link, such as the contact page. 
> These keys
> are
> explained in the Help section of the website, which is referenced by
> <link
> rel="help">. 
> 
> Should we make it a best practice to use numbers as 
> accesskeys? I think
> the
> hijacking problem is a serious one which needs to be addressed. 
> 
> Sometimes, however, I do use letters for accesskeys, especially in
> complex
> forms. I then assign the first letter of each label as an 
> accesskey (and
> make it so that each label has a different first letter). Using CSS, I
> give
> the first letter of the labels a different appearance. These 
> accesskeys
> are
> once again explained in the help section of the website. 
> 
> You could even make it so that a list of accesskeys is 
> inserted into the
> page from CSS if the page is accessed by a non-visual 
> browser. This is a
> CSS2-only feature which is not widely supported yet by assistive
> technologies. (In fact I don't know any that support this but 
> am not too
> familiar with all the AT's out there).
> 
> Yvette Hoitink
> CEO Heritas, Enschede, The Netherlands
> 
>

Received on Friday, 26 September 2003 18:28:57 UTC