Re: Decentralised extensibility idea (ISSUE-41)

Tab Atkins Jr., Fri, 15 Jan 2010 10:09:53 -0700:
> On Fri, Jan 15, 2010 at 9:34 AM, Julian Reschke:
>> Tab Atkins Jr. wrote:
>>> The second isn't really distributed extensibility, because the
>>> browsers have to support each vocab manually.
>> 
>> I disagree that this isn't D.E. You seem to assume the D.E. implies some
>> kind of code-on-demand, which IMHO is not what most people think. Otherwise
>> XML namespaces wouldn't qualify either.
> 
> I'm making a slightly subtler point.  Microdata, RDFa, and even XML
> Namespaces can all be processed with a generic processor that has no
> information about the actual vocabularies being used.  You can just
> say "Find me all the items/triples/namespaced elements on the page",
> and it can do it.  Thus the extensibility granted by these
> technologies is distributed, as no central authority has to recognize
> and bless your vocab for it to be processable by generic tools.
> 
> However, if this proposal requires tools to recognize particular
> @profile values before it can even tell that there *is* data embedded
> in the page, let alone *extract* it, then this does not seem to be
> distributed in any significant way.

I agree with Philip J. in that Toby's DE and MD are similar in this 
point: Both @profile and @itemtype do take URIs. Both MD and DE 
requires you to recognize attributes. However, DE allows UAs to develop 
a deeper support - for the DE spec in use.
-- 
lh

Received on Friday, 15 January 2010 17:32:54 UTC