Re: plotting accessibility

On Thu, Jun 17, 2004 at 09:56:44AM +0100, Roger Gimson wrote:
> Your suggested guidelines (UAs must show the URL, and must be able to 
> edit it etc.) may not be appropriate, for example, for a kiosk UA that 
> only has a touch-screen interface to content from specific servers. The 
> Web imposes no restrictions of the kinds of UA that may be used to 
> access resources, it just defines the meaning of the delivered markup.

Good point... but still a touch-screen interface SHOULD show the URL and
should be able to edit it. It's only polite ! :)

> guidelines, oriented only to current mainstream keyboard-and-screen 
> products, may prevent such evolution.

Showing the URL isn't exactly a keyboard and screen specific thing is
it?

> The work of the Voice Browser activity (http://www.w3.org/Voice/) shows 
> there is demand for such markup.

Yes... but will it be implemented? Used by "authors"? W3C has had quite
a few dud technologies in that respect lately IMO.

> A 'W3C Mobile Profile' conformance mark might be a good idea, at least 
> to help the phone buyer when comparing models.

:)

> CC/PP is only one step towards a more device independent Web. See the 
> latest Working Draft on Content Selection for how this might be 
> presented to web authors:
>   http://www.w3.org/TR/cselection/

Is this server side transformation on XML drugs? XSLT... *cold shiver* I
don't see why authors should be encouraged to write device dependent
content inline.

Am I the only author who does not want "Variability" ?!

Ok, it seems useful for images. But ... why all this trouble?

How about this when you have an image object, the image content server
is able to determine the resolution of the device.

For example:
http://pictures.natalian.org/konica/2004/jun/10/20:22:13/384x512

I have to specify the resolution "384x512" by hand.

BUT what if authors could:

<img src="http://pictures.natalian.org/konica/2004/jun/10/20:22:13/50%" alt="New England style chair" />

My script which serves the image someone detects the environment via the
HTTP headers that my Nokia 3100 has a max resolution of 128x128 and
serves it and 50% of that, so 48x64 (keeping in line with aspect). It's
like two lines of code if I can get the max resolution (hack: perhaps a
UA agent string DB with the max resolution the UA can handle there).

That would solve the nasty problem with bitmap images satisfactorily for
me. Cropping etc. aside. But then I don't see how cselection solves
"centering on interesting point" type cropping either.

Do you really have to invent a syntax for this? And expect the author to
use it?

Received on Thursday, 17 June 2004 05:40:01 UTC