Re: plotting accessibility

On Fri, Jun 11, 2004 at 09:54:46AM +0100, Roger Gimson wrote:
> >It's confusing! d-i should really come up with uniform user agent
> >guidelines. Like UAs must show the url. Must be able to edit it. Must be
> >able to view source. Must be able to handle xxxx size documents. Must be
> >able to report errors. And so forth.
> 
> I strongly disagree. We do not wish to restrict the ability of UA 
> developers to offer different ways of controlling web access. Rather, we 

You won't be restricting them. They could be guidelines they should try
meet. An etiquette. Some developers don't know good manners.

> would like to ensure the delivered markup and metadata is appropriate 
> for the needs of the user (via their chosen UA). So, for example, I 
> should be able to use a voice-only interface to the web if I wish 
> (useful for the visually impaired as well as when driving a car). This 
> might *not* offer the ability to show or edit a URI, but simply to 

Voice and other related modes of interaction IMO is little unrealistic
right now. It is bound to be rubbish. Anyway semi-decent voice
technology is all patented. Lets start with the browser interface.

> invoke links and back/forward. However, the delivered content must 
> clearly be suitable for auditory presentation - either by the origin 
> server offering this as a possible representation for the accessed page, 
> or by some adaptation (in the UA or an intermediary) from text to speech 
> (which requires some sematic markup of the text to do a good job).

No one is going to mark up text for voice.  User agents will have to do
it, with natural language processing.

> >Else where else is this going to be done?
> UA interfaces should not be standardised. It should be a matter of UA 
> providers meeting the variety of needs of the marketplace, and being 
> judged on their relative merits.

Hehe "market place". 

Monopolies rule the web right now. Do you think people will be able to
judge and compare Nokia's XHTML browser? There is no choice once they've
bought the phone.

> >If I pick up a mobile, and use their XHTML browser I would like to know
> >if it's going to work or not. 
> If a UA claims to support XHTML, it should be possible to certify that 
> it actually conforms to the Recommendation (by test suites etc). 
> Incresing emphasis is now being placed on providing test suites and 
> conformance requirements with all W3C Recommendations (see 
> http://www.w3.org/QA/).

Test suites won't happen. Especially automated ones. You need someone to
verify the behaviour of browsers.

As I said earlier, all I would like to see is some mark for authors to
aim for. Right now my mobile XHTML browser doesn't do media queries, has
all sorts of basic UI problems and I still can't figure out the maximum
size of document it can handle. 

"Authoring Techniques for Device Independence" is what many people know
already. What we need is a decent level of browser. I am afraid basic
questions authors are asking are not being answered with CC/PP.

Received on Wednesday, 16 June 2004 03:32:41 UTC