Re: Minutes from ePub session of HTML WG at TPAC

Michael(tm) Smith, Sat, 6 Nov 2010 12:56:37 +0900:
> From: "Michael(tm) Smith" <mike@w3.org>
>> @2010-11-05 21:40 +0100:
>> Julian Reschke, Fri, 05 Nov 2010 10:24:56 +0100:
>>> http://www.w3.org/2010/11/04-html-wg2-minutes.html
>> 
>> Mike, in the minutes, you are quoted for saying:
>> 
>> ]] MikeSmith: Before ? pointed out xpointer, I didn't
>>    know we did registries [[
>> 
>> Then what about the XHTML Vocabulary? We have mentioned it many times.
>> It is also a kind of role registry.
> 
> ...except that it doesn't actually call itself a registry, and 
> doesn't seem to
> have been intended as a registry in the normal sense.

It is a perfect documentation of the the things it pretend to document: 
«… a vocabulary collection utilized by XHTML Family modules and 
document types using XHTML Modularization, including XHTML Role and 
XHTML + RDFa as defined in rdfa-syntax.»

> And it doesn't itself provide any information at all to readers about how
> they might themselves propose new vocabulary items to be added to that page.

The criteria for being listed at that page is that the relations are 
listed in the specifications that it covers. We, in this group, know 
how to create specs.

>> It is not a "registry office"
> 
> I don't think most people would say that it's any kind of registry at all.

We are not looking for a word here.
 
>> though (unlike the xpointer registry), but "documentation registry" -
> 
> I would doubt that what you intend by the term "documentation registry"
> would be clear at all to most people familiar with existing IANA
> registries or similar registries elsewhere.

That page is not 'perfect', but it could be a good start.

>> the actual registries are the specs which the relations and roles on
>> that page are taken from.
>> 
>> http://www.w3.org/1999/xhtml/vocab/
> 
> That seems like an especially atypical use of the term "registry",

The initiated link rel registry and its designated experts ask for link 
relations to be rooted in a spec before they are being taken into the 
link registry. Which is similar to how that page functions.

> and I would think
> that a lot of people might anyway not see how it's relevant to the
> discussion about what would be the best way to implement a proper registry
> mechanism for link relations.

That page will continue to exist, and so it would make sense to 
integrate it - somehow - into this 'proper registry mechanism'. I don't 
know exactly what kind of mechanism Henri will suggest in his CP. But 
it would be nice if for instance approved link relations would appear 
on the Vocab page and if the Vocab page linked to the link relation 
registration form. It would require some co-operation with the owners 
of the Vocab page, of course.
-- 
leif halvard silli

Received on Sunday, 7 November 2010 00:26:15 UTC