W3C home > Mailing lists > Public > spec-prod@w3.org > July to September 2011

Re: Publication of specifications as HTML5

From: Robin Berjon <robin@berjon.com>
Date: Wed, 24 Aug 2011 14:21:10 +0200
Cc: Karl Dubost <karl+w3c@la-grange.net>, Richard Ishida <ishida@w3.org>, spec-prod@w3.org, ayg@aryeh.name, Doug Schepers <schepers@w3.org>, Philippe Le Hegaret <plh@w3.org>
Message-Id: <C3957A14-3369-4389-8E71-91E98B054AFA@berjon.com>
To: liam@w3.org
Hi Liam,

On Aug 24, 2011, at 03:21 , Liam R E Quin wrote:
> On Tue, 2011-08-23 at 15:07 +0200, Robin Berjon wrote:
>> On Aug 22, 2011, at 05:15 , Liam R E Quin wrote:
> [...]
>>> Seems to me a requirement should be that the format issuitable for
>>> archiving.
>> 
>> I strongly agree. I also happen to think that this constitutes a
>> strong endorsement in favour of using HTML5 
> 
> There are two parts. One is technological and I think can easily be
> addressed. For example, HTML documents could contain a link element with
> rel="conformsto" to point to a specific draft, not for validation
> purposes of course, but for archival purposes.

This strikes me as being in practice counter-productive to the stated goal. There is no way to validate that the "conformsto" link is actually the right one. As a result, the first release of the first specification in a group that uses it will possibly use it right and have it point to the right document. But as time passes, if the markup drifts, the editor will have completely forgotten about that detail and it will start pointing to the wrong document with no discernible effect that will cause this problem to be addressed. And since editors starting new documents do that by copying an existing one and changing the content, the mistake will happily proliferate, turning the conformsto link into at best useless information, at worst harmful data.

This is a general architectural problem with data that is supposed to change the interpretation of other content items but can go out of sync with no author-detectable problem. It is a given that it will go stale.

> The second is philosophical. I'm actually 100% OK with using HTML 5 for
> the HTML 5 specification itself. Once HTML 5 is a Rec I'm OK with using
> it for other things too. I'm also OK with using HTML 5 for drafts that
> are moving forward, with the understanding that they have a dependency
> on the HTML 5 Rec.  I wouldn't want an *unrelated* spec to be published
> as a Recommendation right now, today, in HTML 5, just as I didn't want
> RDFa to be used in Recs before RDFa was itself a Rec. This is because,
> if we ask other people to wait for Recommendation before they use a
> standard (which is what "Recommendation" means - we now recommend that
> you use this) then we should wait ourselves.

I may have missed it but I don't think that anyone is suggesting that all new drafts should be forced to use HTML5, simply that it be allowed in pubrules. I am not sure that I completely agree with your philosophical statement, but given that I don't think there are that many groups in the Interaction domain (plus the Ubiquitous Web Applications Activity, which I always forget is not in Interaction) that don't have a dependency direct or indirect on HTML5 I won't strongly object since it will make no practical difference for the groups I'm in :)

-- 
Robin Berjon - http://berjon.com/ - @robinberjon
Received on Wednesday, 24 August 2011 12:21:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:19:18 GMT