Atom discussion, New York City, 18 May 2004

bt bryan thompson, Hicks & Assoc.
bw bob wyman, pubsub.com
db david beckett, ILRT, Univ. of Bristol
mm matt may, W3C
gd glen daniels, Sonic Software
ba ben adida, Creative Commons
em eric miller, W3C
sr sam ruby, IBM
do david orchard, BEA Systems
rs robert sayre
kd karl dubost, W3C
mf max froumentin, W3C


em: Trying to figure if there's a good match. There's a lot of overhead. The overarching aspect of W3C as an organization may not be right for the community. I wonder whether Atom and W3C is a good match. One way to figure that out is to ask about Atom's interest. Rubber stamp, marketing issue, etc.
sr: We need more process. We've been open, and need to have a decision-making process to make the hard decisions. I'm not concerned about RDF. If it's a good idea, fine. If we decided W3C, when would we get to start work?
em: We'd have a WG proposal, suggest discussing at the 2 June meeting.
ACTION: em Add proposal to Atom meeting agenda
em: From there, it goes to the team, and then the membership for approval. We've already done a lot of the work with the team, it's already effectively signed off on. We would need a staff contact and a chair.
sr: Staff contact?
em: Team member who works with the chair, navigates the process, coordinates with other groups
do: Hugo Haas writes proposals for WSDL WG. He's an added resource.
em: We have horizontal groups like i18n and accessibility who review for interop etc.
em: Regardless of where Atom goes, i18n and acc. are very important to W3C, and we will contribute in any case. If this goes to IETF, liaisons with this will be harder to set up. Inside the W3C, our liaison and coordinations are more effective.
em: Chairs and W3C Advisory Committee get 4 weeks to review a WG Charter. The feedback is evaluated, and the Director makes a decision. At this point, the group is live. Then a call for participation goes out. Everybody joins up and shakes hands and talks about what their goals are. Ultimately, you're going to get new people.
kd: People who join commit to a level of work (hrs/wk).
bt: How long has Atom been around?
sr: 11 months. We thought about rolling it back into RSS, but that turned out to be impractical. So we moved forward.
bt: What is your sense of community expectation? 6 months, 1 year, 2 years?
sr: Group expectations may not be reasonable (this August). I'd like the first half of next year. I wouldn't be excited about 2007-2008...
gd: Could float another draft.
sr: I don't want to see multiple specs come out. Atom 0.3 has been stable for 6 months.
do: W3C Advisory Committee has been talking about fast-tracking. First WD could be Last Call. That may be reasonable given the amount of work done.
em: Would need stable spec, stable test cases, 2 interoperable implementations of each part of the spec, and that will move things forward.
sr: That will happen. But you won't commit to exit timing?
em: Can't do it.
sr: Can't create a WG before end of June?
em: No.
sr: Possible to exit in first half of 2005?
gd: Minimum last call length?
em: 4 weeks.
em: You'd have a 0-length Candidate Rec. Then to Proposed Rec.
em: PR is 4 weeks.
bt: Is it worth it? If you move this fast, those interactions that are being sold as the value of W3C aren't going to happen.
em: Yes, there's a caveat. Question is, can it be done in a way that syncs with the specs?
do: Are you optimizing for speed, or for functionality? Take your pick.
bw: That's a false choice.
do: Sometimes it takes time for people to do reviews. Those times are set to create certain review cycles.
bw: For us to get the kind of benefits you offer, how long would you expect it to take?
kd: There is a deadline for comments. Usually, if they're missed, it's over. If there are major issues that arise, they will need to be dealt with. Most difficult part is to answer the comments.
bw: I know it's hard to estimate. I just want a gut feel.
em: My gut feel is 12 months to Rec.
sr: July/August 2005
em: There are some coordination points (e.g., RDF in XHTML) that will need to be made.
em: Often, there can be roadblocks.
do: All you need is one or two things where you don't have agreement.
bw: Concerned about the RSS vs. Atom battle. There are W3C members who could join and get a vote who are in the RSS camp. Microsoft guys keep talking about RSS and trashing Atom. If we're in the IETF, and have RSS vs. Atom, chair can say, fine, we have rough consensus on what Atom should be. You can go do RSS. In W3C, voting no because Atom shouldn't exist. 
em: Vote you're talking about happens in AC, not in WG. And they approved the creation of the WG. IETF example you cite is not exactly the case. Chair's prerogative can be revoked at any level. I was in the URI committee for 6 years, and new people would come in with the same issues. Without a process, you're out of luck.
do: Process for decision-making is one of the things I like about W3C. If a vote happens, and it's 7-3, and the 3 are pissed they lost, there is an appeal process, and the minority opinion is filed in the process for the spec.
em: At the Director's decision meeting, that minority opinion gets evaluated.
sr: Only thing I object to is someone who says "RSS 2.0 as is must be rubber-stamped."
em: And we wouldn't rubber-stamp Atom.
mm: We're interested because we want Atom to be a coherent part of the whole Web architecture.
bw: There are many standards in W3C where they would work with. But there are many things in IETF in which W3C appears not to have shown interest that Atom should work with. For instance, with push, you need a continuous connection. But a plethora of choices in IETF. From my point of view, would be better to work with Jabber/XMPP people, etc.
sr: Nobody's convinced me they can't do both.
bw: If W3C picks up Atom, will they deal with peripheral issues like Atom over push connection?
gd: That's up to the group.
do: You would want something like that in the charter.
em: If we go down that path, the 12 months gets pushed back.
do: You can say v.1 doesn't support push, and 1.1 does, etc.
bw: Suggest that to do that piece of work opens up a Pandora's box. There are other things that come along with it. In IETF, someone will give me a WG if I yell loud enough. But if other issues arise?
em: The issue you're talking about, we haven't considered. That's not a level of expertise we've been focusing on.
kd: Have you discussed on the Wiki?
bw: Not yet. Haven't wanted to derail things.
bw: To make things a bit worse, there are issues about Web infrastructure, like ping over XML-RPC. It's necessary as a part of the ecosystem. The right thing for Atom is to do this.
kd: Charter has deliverables. Nothing forbids starting a new WG after to do more work in the area.
bw: If I say we should add this stuff to the charter on 2 June, and 4 other spec efforts, I'll be laughed out of the room.
do: Do you want these things addressed in a WG, or not?
rs: There are implementations for the pull stuff, and not for the push.
sr: Start bringing this stuff up. I'd like to discuss this on 2 June. I'm excited by this possibility.
bw: It could be a serious delay.
sr: Yes.
bw: There are 17 different eventing specs coming out, 4 or 5 strong candidates in IETF.
kd: Why hasn't this been raised yet? Why is this an issue now for W3C vs. IETF?
gd: And why would push have to be in the spec?
bw: I'm suggesting if people want to be responsible for the spec, they should take on the responsibility for the ecosystem.
em: The most successful efforts I've been involved with are well-scoped and small. Your app is larger than the blogging community. This very much gets back to early library architecture issues like privacy. I'm reluctant to put this into any one charter.
bw: I'm not proposing that. I'm saying if W3C wants this, will you take responsibility for the whole bag? I'm not worried about IETF.
em: I can't answer this for W3C membership. If there is a snag, and the AC says, we don't want this, which is a possibility...
sr: More concerned about AC up front saying we don't want this. As far as exit date, realistically, our current state is blissfully ignorant. We're not aware of any major problems. If at the end of this year, we are still unaware of major issues, I'd like to go for PR at 1st qtr of next year. Jabber, I think it's a matter of weeks. RDF, a matter of weeks. XHTML, I don't know. If at the end of the year, we discover there are real problems, I'm _still_ happy. That's an excellent scenario.
em: I consider that a healthy sign.
sr: At least have the confidence that our things have been reviewed. I'm excited about PubSub, even though it could derail us. Lots of people are saying RSS won't scale. Somebody is going to say I told you so.
bw: Werner Vogels at Cornell has charted it out. We're at the knee of the curve. I don't think we have 2 years.
sr: I have had major media people who say, until you solve this, I'm not in.
bw: However good the spec is, unless we deal with the bag issues, it won't matter. There are fundamental flaws in the current architecture.
gd: That doesn't mean you need to boil the ocean immediately.
bw: Whoever takes this will need to boil the ocean someday.
bt: Should discuss this with AC to show the value.
em: I've been talking with them and asking what they think. The answers have not been no.
bw: There is a RESTafarian prejudice in W3C.
sr: Tim Bray a year ago would have been against this because RDF people were too strong.
em: In the past 6 months, RDF support has gotten stronger. At the same time, I've talked with Tim Berners-Lee, and he has said this will not be done by fiat.
do: For better or worse, there's a perception of two different camps. I have been the recipient on the Web Services side of ideas we thought were not the best approach. Coupling WS Activity to Semantic Web. AC voted no. Expressing SOAP data model as RDF data model. That evaporated. Expressing WSDL in RDF. Pro-WS group has concern that there will be a push of technology. Concern should be addressed more strongly from W3C side.
em: I hope my statements have started to address that. Charter should address that more. I believe RDF is the right approach as I hear more. But if the decision can't be made in the WG, it will not be made by fiat.
kd: I agree with Eric, but I trust the WG community more. That's the important part.
bt: I'd be shocked if we couldn't get member companies in AC to vote for this.
sr: I work for a big company. So does Tim Bray.
do: I expect that if the vote is a no, it would be a resource-allocation counterargument.
em: That's a fair concern.
bt: But could you respond to that?
em: We wouldn't ask if we didn't have that response.
sr: I understand the delay in responding to us. It was so you could fact-check. I think that's a good thing.
do: Advantage is the community of people. W3C is a mixture of vendor-driven and team-driven. It doesn't have as much of a grassroots. There's a perception that it would be difficult to get grassroots into W3C.
kd: In QA WG, we had lots of support from the community, including the Interest Group. Same in WAI.
mm: We have a number of WGs in WAI that are principally non-members. That's worked for several years.
do: Case studies of these would be good.
em: SW Interest Group has a lot of sway with the WGs.
bt: How large would the WG be?
sr: Active editors at any time are 2 or 3. Active submitters of proposals in Atom are maybe a couple dozen.
bt: How would that interact with W3C process?
em: Good standing status is chair's prerogative.
mm: WCAG WG has 100 people on the list who are "active" on mailing list, but 2 dozen in good standing.
sr: We get a lot of posts who say "I think that Atom should x".
mm: That's an Interest Group post. W3C WG people know that if they say "somebody should x," they just took an action item.
em: IG relation with WG is very symbiotic. Issues that get raised in IG carry over.
sr: I was chair of ECMAScript, and have chaired other groups. I know how this works. You have to encourage good work.
do: There's the Zakim phone bridge at W3C, which is excellent for coordinating meetings. The tools are the best.
em: Telecon bridges, IRC bots, process control, validating systems, etc, to keep things flowing.
do: Being able to queue up over IRC, etc., little things like this make meetings much easier.
bt: Is there another home outside SW Activity that might be better?
em: Good question.
do: My first reaction when I hear that is, "Oh, that's going to be RDF." That's just my perception.
em: It's difficult to argue with that.
mm: TimBL looked me in the eye and said, "it doesn't have to be RDF to go to Rec."
sr: I'm at peace with Eric.
rs: Process would be the same anywhere. Web Services I think would be equally bad.
bt: I think it would be a hideous match with WS.
rs: There's really no good answer to that in W3C.
db: SW IG has been open for years. It's been great.
bt: Is this Technology and Society Domain? What else is under T&S?
em: Security, Privacy... SemWeb is the only one that makes sense.
bt: Is Atom better suited as an Activity than as a WG?
em: It's worth considering. Starts shooting ourselves in the foot for timing of deliverables.
sr: I think it's unlikely to happen. But it could move over time.
do: XML Protocol moved under WS Activity.
em: SW Activity closes Jan 2006. If the membership decides the resources should go away, they go away. We wouldn't charter an Atom WG past Jan 2006. It would need to recharter after that.
db: What next?
sr: Post minutes to atom-syntax?
sr: I would like to see more pull from W3C. Work on W3cCharter.
sr: IETF is a black hole. I'm not thrilled with them. But it has an external perception of being open.
bw: I think the statement that W3C's patent policy is RF is disingenuous.
mm: It would require an Act of Congress to add RAND terms.
bw: IETF and W3C are in the same situation.
mm: The RAND policy hawks left the consortium because they realized how difficult it would be.
do: You have to account for the perception of bias toward RF.
bt: It's a perception that accounts for the inclusion of these technologies. I don't know about IETF, but at W3C, you'll find a lot of lateral motion to accommodate an RF solution.
em: The way that this proposal has been handled was not as I would have liked.
bt: My experience with WGs is not that people say "I'm with x company, I'm with y."