Re: PROPOSAL: Add prefixes to RDFa Core initial context

On Jul 19, 2013, at 11:20 PM, Dan Brickley <danbri@danbri.org> wrote:

> On 29 March 2013 12:02, Stéphane Corlosquet <scorlosquet@gmail.com> wrote:
>> On Fri, Mar 29, 2013 at 7:27 AM, Ivan Herman <ivan@w3.org> wrote:
>>> Done. See also
>>> http://www.w3.org/blog/SW/2013/03/29/new-prefixes-added-to-the-rdfa-core-initial-context/
>>> 
>>> I have put up a G+ item (cc-d it to some of you), and tweeted it; it would
>>> be good to beat the bushes for implementers. Is there a way to send out a
>>> reference through rdfa.info?
>>> 
>>> pyRdfa has the new prefixes already. Is it necessary to extend the test
>>> cases?
>> 
>> 
>> A testing for these prefixes would be a good way to make sure all
>> implementations are up to date and have these new prefixes. I'll work on
>> such new test.
> 
> We are updating our implementation. As part of that, I've been asked
> what the plans are for test cases. Did one ever get added?

Part of this question is if the test suite should continue to change after the specifications are released. I certainly see the advantage of having tests that verify that default prefixes, terms and vocabulary definitions are properly handled by a processor.

> Related question: does W3C state its intentions anywhere regarding
> management of the Initial Context? Masahide Kanzaki also raised this
> question when I saw him in Tokyo recently: the context is quite an
> integral part of any modern RDFa parser. It would help if parser
> writers had some good mechanisms for getting notified of changes.

It certainly would. I believe the intention was not to update it too frequently, but I don't believe there's an announcement mechanism; there should be.

> Perhaps RSS-like feeds from the Initial Context page. It's a pity
> http://www.w3.org/2000/08/w3c-synd/ doesn't seem to work any more. Or
> optional low-traffic mailing list.  Or 'what's new' markup in the
> page.
> 
> Currently the page is signed by Ivan, but is inconsistent: "Updated:
> 2012-11-21 $Date: 2013-03-29 10:51:55 $"
> 
> i.e. it shows that something happened in March but doesn't say what.
> The closest I can find to a public history is
> http://web.archive.org/web/20130407124329*/http://www.w3.org/2011/rdfa-context/rdfa-1.1.html
> (although I know it is in CVS behind the scenes). Similarly with
> tests, if there is a new relevant test, that would be great to know
> about in some systematic way.

I agree that including a history of changes to the document, at least in a human readable form is something we should do.

> I guess the main thing is to be clear how often it is likely to be
> updated. Maybe we can just say to check back 'every n months'; 3? 6?
> And that the goal is to minimise improvements.

About every 6 months is I believe the frequency we were looking for, but I'm not sure that's documented anywhere.

Gregg

> (I'm thinking about similar issues for schema.org currently fwiw)
> 
> Dan

Received on Monday, 22 July 2013 05:04:04 UTC