RE: Comments on: W3C Process2015

From: Phillips, Addison []
Sent: Monday, April 06, 2015 1:04 PM
To: Stephen Zilles;
Subject: RE: Comments on: W3C Process2015

Steve wrote:

“A Candidate Recommendation is a document that satisfies the Working Group's technical requirements, has already received wide review, and has been reviewed by W3C’s horizontal groups.”
[SZ] What you are asking for already exists within Process 2014, but it may not be as obvious as you would like. The drafters of Process 2014 were and continue to be concerned about facilitating Horizontal Reviews. In that context several things have been done. The first is that dependencies between a Working Group and various Horizontal Groups MUST be identified in that Working Group’s Charter.

I don’t think using the Charter is realistic or all that helpful to the Internationalization WG. It suggests that there are some charters where we should not appear as a dependency (which I agree with) and thus not be given a chance to do a review (which I don’t).
[SZ] I think you are misinterpreting what a charter dependency is saying. The intent of putting in a charter dependency is to put the onus of insuring necessary (horizontal) reviews are done on the Working Group. Charters are supposed to list intended deliverables and have a scope that is sufficient for Members to decide to join the Working Group and make a patent commitment. Reviewing these should give you a hint if you think a review will be necessary. That said, there is nothing about the lack of a dependency prescribing a review; it just puts the onus on the Horizontal Group to find the document and review it. The W3C established an announcement list,<>, to allow Working Groups to announce drafts for review. Unfortunately, the use of this list has not been well socialized so it is a bit deficient currently. But, if it becomes popular, it will be a good place to check for drafts worth reviewing (versus ones that are only minor updates).

Frequently the organizers of new Working Groups don’t recognize or identify internationalization issues at charter time because what they are developing doesn’t present high-level Internationalization issues. But at the same time, nearly any Specification can have internationalization (or other horizontal group) considerations buried inside it.
[SZ] That is the point of having the charter reviewed by the Horizontal Groups. To get input on likely problems at charter creation time not later when it is too late. Your Horizontal Group is making the decision as to whether to request a dependency.
In the cases where there is no formal dependency in the Charter between the horizontal WG and the spec developers, the responsibility of checking the Spec falls onto the horizontal team, and, unless horizontal teams are a required part of “wide review”, the Process isn’t helping horizontal teams do a good job.

One possibility is that we’re saying that every Charter includes every one (or at least most) of the horizontal groups, but this seems to run against the point of calling out dependencies in the Charter.
[SZ] No, as noted above, the point is to identify things that the Working Group should be taking account of in creating their specifications.
Or it means that we’re okay with “stuff slipping through” because a document with no explicit internationalization (or whatever) needs at charter time ends up having such an (unidentified) issue later.

I agree with Chaals that FPWD is a good time to start the review process.
[SZ] I agree as well and FPWD’s do get announced on the Public webpages so this should be easy to track.
Reviewing at LCWD is not necessarily a great experience: identifying I18N problems when the WG thinks they are nearly done--when code has been written and compromises have been made--is a tough sell.

While the changes I’m requesting are generally editorial to the process, I do still think changes are needed. Most specifically, it should be possible for the Director to know that “wide review” included any and all horizontal reviews that turned out to be needed, even if the horizontal group was indifferent (“we don’t see a need to review this document”). I think that means that horizontal review is an explicit part of “wide review” rather than implied by (possible) inclusion in the Charter and, ideally, that we find some way of knowing that the horizontal groups have been reached.
[SZ] The point of using Charter dependencies was to put some structure into who had the responsibility to reach out and when. Wide Review is a broad topic and it must be satisfied by CR time, but that, as noted above, is too late. The goal of putting the requirement in the Charter was to get the Working Group to contact the Horizontal Groups it has dependencies on and work out a plan for doing reviews at an appropriate time. For some groups that may be FPWD for others it may be a bit later in the maturity process. The important thing is for the Working Group (or its Chairs) to have a conversation with the Horizontal Groups to decide what makes sense for them. This is not something that should go into the process.


Addison Phillips
Chair (W3C I18N WG)

Internationalization is not a feature.
It is an architecture.

Received on Monday, 6 April 2015 22:53:00 UTC