See also: IRC log
<koaliie> Process chapter 7 draft... [chaals, 3-Mar]
SteveZ: Chaals has been extremely busy,
responding to a set of comments
... we reached a document that the TF would like to forward to the AC as a
second Last Call
... the target would be to have a document approved by the AC Meeting in
June
SteveZ: Ralph sent comments, I send editorial
comments too
... we reached a point where comments are largely editorial
... Let's see if we can get a resolution to send this as a final Last Call to
the AC within a week or so
... chaals do you want to give an overview of what happened since our last
meeting?
chaals: We have reinstated a phase called Proposed Recommendation to clarify the transition process from CR to Rec.
... I'm assuming people have looked at the document, or don't care
... if you don't care, please, don't send editorial comments next week :)
... We talked about reinstating PR at the last teleconference
... I don't thing we've made substantive changes other than that.
... Modulo editorial tweaks, the draft is stable.
SteveZ: Are there issues that people would like to bring up for discussion?
Chris: I had a question about CR
... CR signals the AC should review, but not until PR is a deadline set?
... What's the goal?
chaals: The WG should have but may not have
collected implementations experience
... If they haven't done before, they're stuck in CR until they're done
... That's why there is no fixed timeline
chris: Director has to approve the transition anyway, but @@@
Jeff: Once you get to PR you can no longer drop any features
chris: I have no problem with that process; it
might be good to tweak it to say clearly what the goals are
... I had to read between the lines to understand the aspect of implementation
experience
steveZ: And the IPR review
... you need to complete that
Ann: I liked Chris' point about wanting
explanation about each step, and why one moves to the next step. So one doesn't
have to 'read between the lines'.
... I read it again yesterday and mostly have grammatical suggestions
Chris: There were reasons for streamlining this and the clearer this is stated, the better.
chaals: We could cross-reference the stages
Chris: The part I got stuck on was what made a
spec from CR to PR, how long that was supposed to take
... Even moving between the maturity levels wasn't obvious
chaals: My suggestion would be: Adding a note about getting implementation experience to the maturity levels about PR
<chaals> ACTION: chaals to Add a note about getting implementation experience to the maturity levels about PR [recorded in http://www.w3.org/2014/03/04-ab-minutes.html#action01]
<trackbot> Created ACTION-55 - Add a note about getting implementation experience to the maturity levels about pr [on Charles McCathie Nevile - due 2014-03-11].
Chris: re: claims of independent implementations, I don't know if this needs to be captured better, because a spec is a spec, and not somebody's code.
chaals: There is a tension and I agree with the
issue
... The current draft says the group needs to convince the director that it
will get independent implementations.
Chris: It's hard to make it a hard requirement.
[Chris gives experience from Audio spec development]
chaals: We did move to try to address it as far
as I could figure out how
... I didn't find any way to make a hard requirement that would be useful
... The key phrase is "The Director will consider"
chris: "Independent" vs "independently developed" are different
steveZ: One way out would be to define the intent
of "independent" (writing from the specification, or something like that) to
clarify the sense
... in a number of groups, this has been interpreted as "not using the same
code base"
chaals: My aim is to keep the process reasonably
simple and limit the number of clarifications
... because more text also makes the document harder to read.
Mike: Both Chris' comments to me mean that we did
a little more work before we take this to the AC
... What we need is not a sentence or two, but a paragraph that gives the
philosophy
... of what we are trying to achieve
... We can't specify everything in great detail
... and the Director is going to make a judgement
chaals: There is a paragraph already
Mike: I'm worried we might have this conversation
again and again with the AC
... If we give the big picture of the process goal, the role of the Director,
etc.
SteveH: in WSI we ran into the same thing
... independent interoperable etc. We left it to the Board to vote
... some people would interpret the whole open source issue
<inserted> Ralph: and judging whether it was necessary to look at the [open] sources to interpret the spec
chris: That's exactly my concern. And I wonder if we are capturing this well enough. I'm hearing this is captured.
SteveH: We left it with these words and left it
to those who decide to make a decision
... We voted and in this case the Director makes a decision
... People complain if this is a piece of the same code
... And people should complain if this is the case.
SteveZ: Either we leaves things are is (as chaals
said and SteveH said mentioning WSI) or we write about the intent and
clarify
... My preference would be the latter
Mike: Every body is sharing source code, this is the reality. What is exactly the role of the Director? I can't find a general overview of what he's trying to ensure.
chaals: There's a section specifically on what the Director is trying to ensure (7.2.4). I'd be concerned with over-constraining
[strawpoll]
SteveZ: 4 for as is
Ralph: I suggest we give more weight to the thoughts those who were in the TF
<tantek> how about asking for objections?
SteveZ: Anything more needed in 7.2.4?
Mike: Earlier than 7.2.4
Jeff: We started with a list of major changes.
The philosophy of what we are trying to do is lacking.
... I'd suggest an additional paragraph
chaals: The vision of the process revision is an artifact that doesn't make a difference to the people who are going to follow it.
Jim: There is ambiguity as Chris pointed out.
... if in 7.2.4 we said "how independent are the specifications [...]" would
that help?
SteveZ: The other part of the strawpoll: are
there people who feel we need to do more?
... 2
Jeff: In 7.2.4
Mike: 7.2.4 is generally OK
... but I like Jim's suggestion
... I know you want to declare victory, but I think we're going to have this
all over in the AC.
Chris: We just need concrete suggestions
SteveZ: Adding one more bullet point about "how independent are the implementations"
Jeff: It's likely to create more confusion
Jim: In the 3rd bullet, change to ask how independent the implementations are.
Mike: In the real world implementors will copy the code in polyfills even before the work is brought to W3C
Jim: "How independent are the interoperable implementations" ?
chris: My concern is in making "independent" a spectrum
<AnnB_> suggestion: "Are the implementations adequately independent?"
Chris: "How independent are they and how independently were they developed" ?
SteveZ: the TF chair would like to take Chris'
suggestion to modify 2nd bullet
... I believe there is consensus
<cwilso_> are there independent interoperable implementations of the current specification, and were they developed independently from each other?
RESOLUTION: Update the 2nd bullet with "are there independent interoperable implementations of the current specification, and were they developed independently from each other?"
Jeff: I was trying to find the whole question of how and where and when we're going to introduce the process. Is that in this chapter?
SteveZ: No, and it should not be
Jeff: When we ask for the AC approval, we need a cover letter
<scribe> ACTION: SteveZ to include the resolution of issue-39 in the cover letter [recorded in http://www.w3.org/2014/03/04-ab-minutes.html#action02]
<trackbot> Created ACTION-56 - Include the resolution of issue-39 in the cover letter [on Steve Zilles - due 2014-03-11].
Jim: The vast majority of the AC may not read it. Will you be sent to CR stage?
chaals: I presume the Director reads the AC comments. early AC feedback is likely to come up from AC reps who care (whether about getting it done or stopping it). And the start of Patent review should signal time to look at it for interested parties.
Mike: The TF overruled the issue I'm going to
bring up.
... We need to say a bit more in 7.1 about what happened to last call and to
address the concerns of people for supergroups.
... My suggestion is to put an additional sentence or two at the end to say
that those who find LC valuable can put it in their WG charter that they will
use it in their process even though the process doesn't require it.
... The other suggestion: In order to help at the beginning of the document,
state the goals, acknowledge that more mechanisms may be needed.
SteveZ: The end of 7.1 says that already
Mike: We know what that means. Does the AC know that that means?
SteveZ: The definition of wide review says that LC isn't necessarily sufficient.
Mike: Fair enough, I wanted to see if the whole AB wanted to raise this as an issue.
SteveZ: Does anyone (beside Mike) feel we need to add something to the document?
[none]
chaals: My goal is to say this is good enough, it
will work, give it to the AC
... The longer we take to give it, the less agile we are.
... there is diminishing returns on tweaking the text; this is why I asked for
concrete suggestions
<Zakim> SteveZ, you wanted to say, do we have a PER
SteveZ: In my review, I noted that in the Edited
Rec section there was still a reference to Rec that we dropped
... It wasn't clear to me whether there was exclusion opportunity
... I was a little bit concerned about referring to terminology that we
dropped above
chaals: It may be a Copy and paste mistake
... the fundamental question is whether you have to publish a proposed rec or
you ask the Director to publish a Rec
... 7.7.2 should point that the process of publishing an edited Rec means you
go through Proposed (edited) Rec...
Jeff: and change the bullet(s) underneath
Ralph: Be careful about the fist bullet
underneath
... This is the Director who publishes, not the WG.
chaals: Yes Ralph, I agree.
Ralph: Something just came up in an internal
conversation
... the question is when we're rescinding a W3C Rec
<Ralph> Rescinding a Recommendation
Ralph: Text in the current process says that
there is a publication of a rescinded Red announced that encourages the Team to
update the status of the thing that was rescinded
... to note this is no longer a Rec
... in my reading, there is no language in the current proposal to suggest the
Team should update a W3C Rec is rescinded
chaals: I agree, and there should be; obvious oversight.
Ralph: The current process doesn't suggest how it
should be done, just that it should be done
... The current proposal doesn't even have a suggestion
SteveZ: Accepted
chaals: WFM
Ralph: I'll provide specific language, if you
wish.
... Do you need to see specific language? [no]
Jeff: Let's assume that HTML5 points to Dom4 and years later we rescind Dom4
Ralph: Awkward state for W3C to have normative references and we should fix this, but adding language to the process is not necessary
<Zakim> Ralph, you wanted to reply to Jeff
<Ralph> Jeff: perhaps language that suggests that Working Groups who are responsible for such Recommendations review them
Jim: I agree with Jeff
SteveZ: Proposal is to add a bullet to 7.9
Jeff: Or add a new paragraph
SteveZ: the proposal would be to add "WGs or organizations should review their specifications"
chaals: I object to the words and the intent of
the proposal
... There's no way that we can be sure that real action is taken, as Ralph
said.
... Groups are developing specs under the patent policy, under their charters,
etc.
... if they're not solving a problem, the Process isn't going to be able to
solve it either.
[Jeff gives an example]
Jeff: What happens to normative references if the process doesn't say a thing about changing but only about the future?
chaals: you can update with a new recommendation, but a stable rec can't be changed; it's part of the guarantee of stability
Jeff: True.
... That's the reason why I proposed mild language.
Jim: In the case where a group has disappeared, what is the right thing, who has the responsibility for the actions?
Jeff: W3C has
... in deciding what to do about it, W3C would have to determine how bad and
awful the normative references are
... and W3C might do nothing if the group has disappeared
... or convene a new group to fix the situation
Mike: I'm tempted to agree with chaals. On the other hand, if we find spec bugs, we'd have to go back and fix normative references, maybe in that world we want to provision it in the process
chaals: on the one hand, legacy is big. On the
other, who cares? What's the practical impact?
... It's a good idea to clean up
... But as Mike said, solving the purity problem is possibly not worth the
effort.
Jim: If the world were the way you describe, I understand. But as Jeff said in his example, there are implications.
chaals: that's why we ask for wide review
SteveZ: summarizing:
... two views
... Wide review will naturally trigger the right thing
... the other view: It's worth putting a sentence.
... Strawpoll
Steve: Raise your hand (4) if we should add that "WGs or orgs should review their specs"
SteveZ: "Organizations which reference this specification should determine whether any action is necessary on their part."
RESOLUTION: Section 7.9, paragraph 3, add "Organizations which reference this specification should determine whether any action is necessary on their part."
SteveZ: We're ready to forward this document as a
second last call to the AC, with the idea of immediately following this with an
AC review to adopt the document.
... with a cover letter that outlines what we're trying to do with the Chapter
7, and mention issue-39.
Jeff: It seems sensible to me that when chaals has finished his revision he can send it to the AB and give us a week.
SteveZ: I agree, this was my other alternative
Jim: I'd happily rely on chaals
... I have a question on the cover that goes with it
... What about comments outside of chapter 7?
... Are we soliciting comments on the whole document?
SteveZ: The reason for sending the whole document is for people who are concerned that the Chapter 7 doesn't match the whole document
Jeff: We're going only to focus on comments on Chapter 7.
<scribe> ACTION: SteveZ to emphasize that the cover letter should mention comments on Chapter 7 are sought. [recorded in http://www.w3.org/2014/03/04-ab-minutes.html#action03]
<trackbot> Created ACTION-57 - Emphasize that the cover letter should mention comments on chapter 7 are sought. [on Steve Zilles - due 2014-03-11].
SteveH: We need to give the materials to the AC. We're in early March already.
SteveZ: 2 steps left: last call and review. That
will take 2 months.
... The ideal is for the review to end after the June AC meeting so the people
can discuss it productively at the AC meeting
... but that it be opened before the AC meeting
Jeff: I'm not sure I agree that the review should last 2 months.
SteveZ: Last call is 4 weeks and an AC review is another 4 weeks.
Jeff: OK.
... The last time we did not send the full document
SteveZ: What the AC will review will be a full
document as this is what is put forward for adoption
... And for the sake of cohesion with the rest of the process, we're going to
send a full document.
RESOLUTION: Chaals will make the edits based on the discussions and submissions of comments to date, we'll have a seven-day review period by the AB, and then a four-week LC to the AC of the full document.
[unanimous support]
<AnnB_> and lots of kudos to those who worked on this, but especially Chaals!