RE: Spec organizations and prioritization

I think it likely needs to be a combination of both making it easier and more resources.  I think it can be easier without changing the process, if we change some things that are not required by the Process.  The idea of the following is that with a target date for a future snapshot, it could be fairly quick between snapshot and AC review of a Proposed REC -- not something that takes months to start after the snapshot.

1. mark sections that have been through last call and have not had substantive changes since the beginning of last call (where all last call comments and issues have been resolved)
2. mark sections that have 2 widely used implementations that the WG considers compatible (and note where used successfully by content creators without compatibility issues).
3. normative sections depend only on normative sections that are past last call (if not, note what it depends on).
4. for some drafts (like the first version of html5), waive the requirement to have test cases for everything and to test implementations.  Test suites are not required by the process.  It's a good thing, but given the long delay for specs like html5, should be waived just to get an html5 REC out.
5. consider publishing sections that are not ready for REC as informative/experimental rather than holding up REC, so readers can at least see what is being considered (e.g. html5 tags that don't have 2 widely used implementations get published as experimental.)  Include a Warning that informative/experimental sections could change or be dropped and do not fall under W3C patent policy.

At a target date, a snapshot would be taken.  Any section that didn't meet the requirements above would automatically convert to being an Experimental/Informative section.  If the WG agreed, W3C staff would check that the above holds for all normative sections.  If it did, it would go into Candidate REC and Proposed Rec simultaneously with the AC review and director's decision.


>-----Original Message-----
>From: Anne van Kesteren [mailto:annevk@opera.com]
>Sent: Monday, March 26, 2012 3:24 PM
>To: Jeff Jaffe; Carr, Wayne
>Cc: Marcos Caceres; Dominique Hazael-Massieux; public-w3process; Daniel
>Glazman
>Subject: Re: Spec organizations and prioritization
>
>On Mon, 26 Mar 2012 22:13:53 +0100, Carr, Wayne <wayne.carr@intel.com>
>wrote:
>> No one is arguing that web standards are final and immutable or that
>> we're all going to finish this soon and walk away.  But it also isn't
>> that every morning we throw away our Web Browsers and replace them
>> with something completely different. Everything is not in flux.
>>
>> In W3C, the output of a Working Group is a Recommendation.  Nothing
>> about it is final other than that particular dated version is done and
>> the next version will have a different version number and date.  At
>> that point in time, it's what is agreed on after careful review.  For
>> something like HTML, I think it would be better for a Recommendation
>> to come out every year (rather than have the last one be 1999).
>> Nothing in the process implies that the technology freezes or that
>> there will never be an updated version.  Nothing in it implies it
>> wouldn't be a good idea to implement drafts that are later than the
>> recommendation if you think they're worth using.
>
>This is indeed very reasonable.
>
>As I mentioned before, we are constrained on resources so either 1) publishing a
>Recommendation each year needs to become trivial, or 2) we need to find
>resources that fork an editor's draft every so often and take it along the required
>steps. Finding resources is something we have been unable to do for the past five
>or so years. Dominique seems to think it is a real possibility again and I am happy
>for him to try, but I think as a group we should focus on 1.
>
>
>--
>Anne van Kesteren
>http://annevankesteren.nl/

Received on Monday, 26 March 2012 23:51:45 UTC