Re: Accesskey again

On 11/27/02 6:36 AM, "Jonny Axelsson" <jax@opera.no> wrote:

> On Wed, 27 Nov 2002 15:17:42 +0100, Steven Pemberton
> <steven.pemberton@cwi.nl> wrote:
> 
>>> On Wed, 27 Nov 2002 13:06:39 +0100, Steven Pemberton
>>> <steven.pemberton@cwi.nl> wrote:
>>> 
>>>> The current proposal is now:
>>>> 
>>>> A user agent SHOULD just focus when an accesskey is used.
>>>> A user agent MUST supply a user option to choose between focus and
>>>> focus+activate.
> 
>> Hmm, first you say you can live with it, then you make changes!
>> 
>> Are you suggesting to make the requirement to switch methods a SHOULD
>> instead of a MUST?
> 
> *I* can live with it, since I believe accesskey to be several mechanisms
> and that the action should be configurable somehow (e.g. with XML Handlers)
> . But you are asking *every* accesskey implementation, even the one that
> currently do focus, to support two different modes and a switching
> mechanism between them. I would suggest saving that for XHTML 2.0. Before
> that, all three implementations, focus, focus+activate, or user
> configurable should be conformant behaviour in my opinion.

Note that this discussion began as an errata for HTML 4.01 - this important
detail was somehow lost.

I assert that for an errata, the additional "MUST" requirement is
inappropriate and out of scope.  I suggest the following:

Errata for HTML4.01:

 A user agent SHOULD just focus when an accesskey is used.
 A user agent MAY supply a user option to choose between focus and
  focus+activate.

Text for XHTML2:

 A user agent SHOULD just focus when an accesskey is used.
 A user agent SHOULD supply a user option to choose between focus and
  focus+activate.

The reason for using SHOULD for the user option instead of MUST is due to
the fact that it may be very inappropriate for certain devices to have to
provide that configuration option.

In fact, there are "standalone" UAs, such as web kiosks that don't have ANY
configuration options at all.

There will also likely be XHTML Basic UAs which do not even connect to the
traditional web, but are instead embedded inside a device and there only for
a special purpose, having nothing to do with the web.

Thus, in general the XHTML spec is the WRONG place user interface
requirements upon the UA.

Tantek

Received on Wednesday, 27 November 2002 11:18:50 UTC