Re: Non-substantive CR and Director's decision

> On Jan 6, 2016, at 15:34 , Chaals McCathie Nevile <chaals@yandex-team.ru> wrote:
> 
> On Wed, 06 Jan 2016 23:22:13 +0300, Carine Bournez <carine@w3.org> wrote:
> 
>> Mostly adding some more background to the questions
>> 
>> On Wed, Jan 06, 2016 at 10:00:22AM -0800, David Singer wrote:
>>> Can I see if I can list the questions?
>>> 
>>> 1) My understanding is that the W3C Process allows the publication of a revised candidate recommendation *without* Director???s approval if there are *no substantive changes*. It would also mean that no call for exclusions are issued as well.  Is that a correct understanding?
> 
> I don't think so.
> 
> If there is a new document, the exclusion clock on it starts, *for the delta*.
> 
>>> 2) if a member excludes a patent about a feature at risk that is removed during the exclusion period (of the original CR publication), might it invalidate the whole exclusion mechanism?
> 
> I don't see how. This is the scenario you mean, right?
> 
> 1. The feature is at risk,
> 2. The patent is excluded,
> 3. The feature is removed from the spec in a revised CR.
> 
> If step 3 doesn't happen, W3C are meant to start a PAG. Step 3 obviates the need for one, or resolves the PAG if there is one.
> 
>>> 3) Is it possible to republish a (editorial only) CR during the exclusion period of the previous one?
> 
> Sure.
> 
>> People do review the "original" CR. They might refer to section numbers
>> for that, same for exclusions. It might be just a practical problem, but
>> if we allow for republication that changes the section numbering during the review/exclusion period, it could cause some trouble.
> 
> Yes, while this is very much about practice rather than formal process, it can be a real issue.

I actually don’t see a problem.  All the documents are stable behind stable URLs; the document that is the Reference Draft of an Exclusion Opportunity is do identified. That’s what Exclusions are filed against, and so any section numbers, quotations, references, must refer to that.

If we can’t work out where that material is after a non-substantive re-organization, I have doubts about whether the re-org was non-substantive.

(e.g. the Excl was filed against 7.3.1.2 but the editorial revision introduced a new 7.2 section, which means that that section is now 7.4.1.2.  We ought to be able to work this out.)

> 
>> Removing a feature at risk may trigger such a section numbering change and we could see an exclusion against the wrong part of the spec.
>> 
>> My take on this:
>> 
>> - Either say editorial CR is after the CfE period (60 days after pub),
>> or
>> - say republishing is allowed during the CfE period, the
>> reference draft would still be the original CR and disclosures can't
>> use section numbers for identification for the portions of the spec covered by the claims.
> 
> I don't think we need to set formal process rules, but I suggest we recommend groups wait 60 days to republish if practicable, or retain the original numbering, identify the changed sections with both a human-readable summary and a link to all the changes, and a 'prominent notice' like the "this is a Working Draft and the Editors' Draft may be more up to date" thing that is sometimes put on specs.

I agree that doing updates to a draft that’s under an exclusion opportunity is maybe unwise, but I don’t see it as formally problematic.

> 
>> For the review comments management, I think it's up to the WG to deal with the practical aspects, as long as they still address the comments received before the editorial republication.
> 
> Yeah, they need to address all comments. End of story…
> 
>>> 4) What is the mechanism for resolving a disagreement about whether an update is ???substantive??? or not (notably, when it is published believing it is not, and someone disagrees)?
>> 
>> Team used to evaluate whether a change during LC was substantive or not.
> 
> That is meant to be covered by the fact that there is a requirement for approval. Where there is nothing controversial, that is routine, but if there is disagreement "someone" can request that before formal approval is given the disagreement is evaluated. In practice, the Team contact should do that routinely if required.

Yes.  If someone feels that a revision was, in fact, substantive, contact the team, and they will discuss, and unless they both agree it’s not substantive, it gets to be considered substantive.  i.e. in cases of doubt or disagreement, we take the more conservative road.  I doubt it’ll ever happen, but we can document it.

> 
>>> are there other questions?
>> 
>> How do we identify an editorial publication, from the document itself?
>> With automated publishing in mind, we need something to make sure it does
>> not get a CfE (we don't want to miss a CfE for a substantive one either).
>> I suppose we don't want automatic publishing of the substantive CRs, since there's a director's decision, but there is no apparent difference between an automated publication and a manual one so we can't rely on
>> that for CfEs.
> 
> cheers
> 
> Chaals
> 
> 
> -- 
> Charles McCathie Nevile - web standards - CTO Office, Yandex
> chaals@yandex-team.ru - - - Find more at http://yandex.com

Dave Singer

singer@mac.com

Received on Thursday, 7 January 2016 18:46:08 UTC