Re: XML namespaces on the Web

On Wed, Nov 18, 2009 at 10:27 AM, Leif Halvard Silli
<xn--mlform-iua@xn--mlform-iua.no> wrote:
> On Wed, 18 Nov 2009 09:31:55 -0600, Shelley Powers wrote:
>> On Wed, Nov 18, 2009 at 9:01 AM, Julian Reschke:
>>> Julian Reschke wrote:
>>>> Lachlan Hunt wrote:
>
>>> Clarifying: I disagree with "...we can all agree on ... they have a need to
>>> implement non-draconian...".
>
>> Forgive my denseness here, but can someone point out the message
>> in this thread that is directly relevant to a possible change
>> in the HTML5 specification?
>>
>> Anne mentioned something about being tasked at TPAC for providing a
>> solution for making namespaces "usable on the web". That seems to
>> exceed this group's charter a bit. Especially when we get into
>> discussions about XML5.
>
> We discuss whether to change text/HTML or not w.r.t. to namespaces. I
> think  Anne's idea with XML5 was: Let namespaces happen in XML. Not in
> text/HTML. Thus his message relates to HTML5 in the sense that if we
> follow his proposal then we will /not/ need to add namespaces to
> text/HTML. I assume that many vendors think that it is easier to add
> recovery to XML than to add namespaces to text/HTML.

Thank you very much for recapping the discussion and reflecting this
back to our work with the HTML WG. The thread makes a lot more sense
now.

I think that since there is no current, organized effort regarding
XML5, we can't depend on some possible, nebulous change happening with
XML in the future, as a way of avoiding dealing with the namespace
issue in HTML5 that we have right now.

>
> Anne proposes making XML more attractive for authors and vendors by
> adding
> recovery as a solution to the "distributed extensibility on the Web"
> problem.

Sure, and if it ever happens at some time in the future, we can
address possible changes to the HTML specification at that time.

The question then is: Would this make Microsoft support XML
> in Internet Explorer? And what about the XML folks? It sounded to me as
> if many of them could accept such a recovery spec, as long as the
> result of the recovery is clearly labeled as "not XML". It is unclear
> to me why you and Julian disagree with that approach.
>

I typically never go down the "What would Microsoft do" path.
Microsoft is a Big Corporation. Tech departments in Big Corporations
typically don't indulge in speculation because they don't want to be
tied down to something some future middle manager won't agree to. That
way they won't get trashed if they don't provide what was seemingly
promised in the past.

As for recoverable XML, I'm not adverse to the concept. And I hope the
XML Core takes a closer look at this, in this thread and in new
efforts.

But whatever happens will most likely happen too late for HTML5. So,
not I don't consider it a viable option for dealing with current HTML5
distributed extensibility, namespace issues and bugs.

> Maciej, OTOH, in his "XML and HTML differences" reply in this
> thread[1], talked about  making text/HTML more attractive, by making it
> more like XML. I believe that with Maciej's approach, then text/HTML
> (with namespaces support) could become the recovery format for XML.
> Opera - and only Opera - already offers users a button to reparse
> malformed XHTML as text/HTML. I don't know if your or Julian finds
> Opera's behaviour acceptable, though ...
>

Oh, I've always been a big fan of Opera's approach. Elegant, helpful,
useful -- well designed, excellent usability.

But one doesn't need to change the XML specification in order to
support this functionality. There's nothing in the XML specification
that states error handling must be useless and punitive, and
completely disregard the environment in which the parsing occurs.


> The important question is what the effect of XML5 would be on HTML5: In
> theory HTML5 would become more of transitional format if user agents
> would support XML5. But the other possible effect of delegating
> namespaces to XML is that text/HTML is left in the same dark as it was
> left in when W3 shifted its focus to XML and XHTML.
>

As far as I know, XML5 is a glimmer in Anne's eye. True? If so, it has
nothing to do with HTML5. Not unless we want to hold up development on
HTML5 for a year or more on what is nothing more than a gamble. No
offense to Anne, and his suggestion.


> [1] http://lists.w3.org/Archives/Public/public-html/2009Nov/0414
> --
> leif halvard silli
>

Shelley

Received on Wednesday, 18 November 2009 16:51:59 UTC