Re: WCAG 1.0 CP 6.1 Considered Harmful

On Wed, 26 Sep 2001, Sean B. Palmer wrote:

> any rate, that is the situation for XML

XML, of itself, is a language for defining markup languages.  It
doesn't inherently have any presentation, so I agree with your
conclusions.

> People have obviously gotten so used to the built
> in styles... but it didn't have to be that way.

But once you have used your mata-language (SGML or XML as the case may
be) to define a markup language, then that language has semantics in
terms of real content.  The semantics of HTML are publicly defined, so
anyone would be free to devise one or more styles for presenting that
markup.

> And it doesn't have to be
> that way for XML;

It can't be that way for XML, at least for the reason I just
indicated.

> > [...] The whole point of decoupling the stylesheet mechanism
> > from the logical markup (as presaged by the HTML2.0
> > specification) was to enable different stylesheets to be applied
> > for different presentation situations and needs.  After all these
> > years, that is finally coming to fruition,
>
> Yes, but I don't see how that is relavent to our discussion.

It was in response to your remark:

| certainly it's a shame that the alternative (stylesheets) were not
| recognized until some time later,

> > Are you by some mischance basing this on the assumption
> > that there would be one single ideal "default" presentation -
> > irrespective of browser or browsing situation?
>
> Blimey, I know that my writing style is at best awful, but I'm suprised
> that anyone could jump to that conclusion :-)

Excuse me, I wasn't attributing that mistaken belief to you - I was
trying to draw you out on whether you were trying to attack that
mistaken belief, in order to better understand what you were saying.

> I'm saying that XML language
> authors should come up with a default stylesheet hint,

Indeed.  There already exist such stylesheets for HTML.

But now you are talking about the authors of the markup language,
rather than the authors of document instances, no?

> UAs may or may not
> implement this, but certainly shouldn't be expected to, and that authors
> can apply any style that they want to their instances.

And the reader can choose to use them, or not, or apply their own, as
they need (or merely "as they please").

> The "author proposes, user disposes" axiom still stands.

Indeed.

On the one hand, if you have used XML to define a new markup that
doesn't have generally-known semantics, then the UA clearly is unable
to offer any useful default presentation in end-user terms.

But if, on the other hand, you are using some (incidentally XML-based)
markup that _does_ have generally-known semantics (let's assume that
others will become popular aside from HTML), then I'm still a bit
confused why you seem to be intending to deny the UA designer the
option to offer some default presentation of those known semantics.

Anyhow, thanks for the clarification.

All the best.

Received on Wednesday, 26 September 2001 07:11:03 UTC