Re: Discussion with Ian and Henri about HTML5+RDFa (part 2/2)

Henri Sivonen wrote:
> 
> When were you meaning to start that thread and how would your preferred
> thread starting time have related to the point in time after which the
> time for finding compromises would have been gone?

I think this comment goes to the core of my issue with how HTML5 is
handling various external work: you seem to think everything should be
discussed within the HTML5 mailing list.

The RDFa mailing list was open for public comment for years, with a
strict W3C process for ensuring that we took into account this public input.

> Furthermore, I
> strongly disagree with the characterization of the over a decade-old
> text/html vs. XML DOM differences as an issue of personal taste.

I apologize that my email gave that impression. The DOM consistency
issue is one I agree with. If one were to go with my personal taste, I
would say "what's the big deal with xmlns in HTML5?" But since there
*is* a path that is backwards compatible using @prefix, then I'm happy
to put my personal taste aside to achieve DOM consistency, which, I
agree with you, is not itself just an issue of taste.

[...]

> Please note, however, that the context of my remarks about following
> your nose was that I thought dereferencing something similar to a fixed
> profile URI in order to discover short name to URI mappings in a
> microformat-ish speculative idea (not RDFa) would be bad for the same
> reason that mostly constant DTD URIs are bad.

And in that case I think it is indeed important to take your point into
account, especially since this is *not* a settled decision. I very much
appreciate this feedback.

[...]

> With CSS, it is super-easy to assess the immediate effect of including
> CSS and whether you value the effects CSS gives you.

And I agree that we need to continue our efforts to provide more rapid
feedback to users as to what semantics they've expressed. Fuzzbot,
SearchMonkey, etc... are all tools that promise to help on this.

But there's no doubt that semantics is harder than presentation. If the
issue you have is that you don't think semantics have a place in HTML at
all, then that's a perfectly fine opinion, though it would be good to
make that clear so we can evaluate your other opinions in that context.
(For example, if I thought adding Math formulas in HTML was a bad idea
in general, it would be best not to consider too seriously my specific
opinions on how to integrate MathML into HTML.)

>> Removing CURIEs is not an option at this point,
> 
> If that's the case, I guess it's pointless for me to try to be proactive
> about coming up with a compromise.

I appreciate your receptiveness to RDFa in recent emails, so I'm happy
to continue the conversation. That said, there's an important question
here. You're asking for a compromise between the existing RDFa standard
and .... what? What is the other argument in this discussion that
requires an existing W3C REC to be compromised? Is there another
standard or component of HTML5 that is breaking?

> Note that the difference in xmlns:foo DOM representation in text/html
> vs. application/xhtml+xml in *implementations* and *deployments* goes
> far further back in time than even the drafting of RDFa.

Again, I apologize that I gave the impression that DOM consistency is
not important. It is. I agree with you.

> Is @prefix not an incompatible change to RDFa--especially if also
> applied to application/xhtml+xml for DOM Consistency?

If we disallowed xmlns in XHTML+RDFa, yes it would be backwards
incompatible. But if we allow it as an alternative in XHTML and as the
only way in HTML, then we've got backwards compatibility for existing
XHTML+RDFa markup, and a path forward for DOM consistency and single
markup across XHTML and HTML.

-Ben

Received on Monday, 26 January 2009 19:21:53 UTC