Re: CDR Framework: Last Call Comments

Hey Ian,

Speaking for myself only ...

On 2/21/06, Ian Hickson <ian@hixie.ch> wrote:
>
> On Tue, 21 Feb 2006, Timur Mehrvarz wrote:
> > >
> > > I fear that CDF is encouraging, or justifying, a "split Web" situation
> > > with multiple profiles, which is directly counter to the
> > > device-independent design of the Web and of W3C's design principles.
> >
> > Why would 'such devices' not interop with existing Web content?
>
> Given a hypothetical profile W that describes the features used by
> existing Web content, any CDF profile P that is not a superset of W will,
> by definition, not include all of W. If implementations implement P, they
> will, therefore, not support all the features used by W, and by extension,
> will not interoperate with the existing Web (which uses W).

FWIW, that doesn't hold in the case when P is defined such that
conforming user agents are required to ignore unknown content.  That's
not quite the case for XHTML, because XHTML Basic (unfortunately)
defers to XHTML 1.0 for UA conformance, and it says that unknown
content must be processed.  In practice though, I would expect XHTML
Basic implementations to ignore unknown content - my implementation
did at least, back in 2000.

On the other hand, while processing unknown XHTML content can result
in a mess on occasion (e.g. nested tables, which XHTML Basic doesn't
support), in general I wouldn't expect it would be terribly disruptive
to the end user for most content.  I admit that's just conjecture
though.

But I guess that's all just a long winded way of saying that this
issue isn't quite as clear cut as you suggest above.

Cheers,

Mark.

Received on Thursday, 23 February 2006 02:36:20 UTC