Social Web Working Group Teleconference
06 Sep 2016
See also: IRC log
- sandro, aaronpk, rhiaro, csarven, eprodrom, cwebber, KevinMarks, tantek
- Summary of Action Items
- Summary of Resolutions
Approval of minutes 2016-08-23
<eprodrom> PROPOSED: approve https://www.w3.org/wiki/Socialwg/2016-08-23-minutes as minutes for 23 Aug 2016 teleconference
eprodrom: please review, pretty productive meeting last time
<tantek> +1 that was a good meeting yes
eprodrom: anyone still reviewing?
<tantek> (despite some of the larger i18n issues being unresolved, both among our specs, and frankly, for W3C as a whole, i18n for JSON strings etc.)
RESOLUTION: approve https://www.w3.org/wiki/Socialwg/2016-08-23-minutes as minutes for 23 Aug 2016 teleconference
<tantek> (aside: I have raised these broader i18n/JSON issues with the AB from a process perspective, proposals without incubation etc.)
Upcoming f2f at TPAC
eprodrom: 2 weeks
<Loqi> Social Web WG Face to Face Meeting in Lisbon (F2F7)
<cwebber2> upcoming so fast omg!
eprodrom: If you have not yet ... is it still possible to register?
tantek: it is, just more expensive
... registration fee doubled on sept 2 and room blocks may have expired as well
sandro: there are defintiely other housing options
<annbass> Oops ... Sorry; late regrets. Taking care of old friend in RI. No way to watch him & also be at meeting. See y'all in Lisbon!
eprodrom: I'm not going to be able to make it, I'll add my regrets
... Company conflict
... But will be remote
... I guess we have one more telecon before it happens
... Any questions?
sandro: We should do agenda planning, at least among chairs
eprodrom any further discussion about the f2f?
eprodrom: Important for us to get an agenda up... make sure we fill that up
... Anything else?
tantek: I do have one concern which is that I think this might be our first f2f with no w3c members signed up to attend not counting chairs
... for actual wg participants
... Which is pretty unusual for a meeting at tpac
... If you're on the call and a w3c member we'd like to know why you haven't signed up
... A concern particularly wrt rechartering
sandro: In my mind that seems *noise* this group, that there is very little industry interest in what we're doing. Unfortunate and would be nice if we could figure out how to change it but has been the case for a long time now
tantek: we did have people at the different f2fs, this is the first time this has happened
sandro: it's on the trajectory, not surprising
tantek: just from a process perspective, makes it much harder to propose rechartering, we used w3c resources paid for by members, so makes it hard to justify continunig the wg when there are no members participating
sandro: we need 20 members to charter a new group
... 10 or 15 actually participating, typically. We're a long way from that
tantek: we might discuss at the f2f, we have chartering on the agenda
sandro: we talked about doing some sort of session at the unconference thing on wednesday
sandro: maybe plan a little more about that, there's a wiki page, we can write a pitch for that now
tantek: encouraged to write up proposals for that
eprodrom: you mean this week, not right now?
tantek: any time between now and the end of tpac
sandro: over the next 2 weeks
tantek: when there is insufficient member interest the w3c typically recommends becoming a CG or IG which don't take up as much resources, but allows anyone who is interested to continue having some w3c related forum to discuss these things
... We'll look at all those options
... If anyone has any other proposals or brainstorms for possibilities we're open to hearing different proposals
... just want to keep that in people's minds as we get close to charter end
AS2 next steps
<tantek> sorry for the charter sidetrack - I wanted to make sure people are thinking about it, because end of charter also affects how we decide to move forward or not with our documents in progress.
eprodrom: I'm the coeditor but I don't think there's much of a need for me to step aside as chair
... AS2 core and vocab went live as CR this morning
eprodrom: pretty exciting
eprodrom: We had agreed 2 weeks ago to push forward with the CR. Amy and sandro did a lot of work last week getting the document in shape for publication
sandro: entirely Amy
<Loqi> rhiaro has 235 karma (125 in this channel)
<tantek> rhiaro++ thank you!
<Loqi> rhiaro has 236 karma (126 in this channel)
eprodrom: I guess my question is what happens next with AS2?
sandro: basically we need to get out the word to anybody who might implement
... tell them now is the time to start implementing
... switch over from as1 to as2
... help them with their testing
... and gather results of testing
sandro: then we can look at results, and hopefully go to rec. Or possibly do revisions which we don't really have time to do, but maybe some quick one
tantek: Sandro covered most of it. 3 basic things - outreach to developer community, now is the time to update implementations. this is ready fo rlive development and use
... Please file issues asap
... The second is substantitive issues are always likely to come up during CR so we need to address those
... They may require normative substantive changes. Usually a good sign, means that developers are digging into the spec and finding things that none of us found
... At som epoint we have to decide if we've handled all of those, and potentially do a second CR
... Third thing is that we need to complete the test suite, or declare that we have, which I don't think we have, but basically say everything in the spec we have at est for and we need implementations to do those tests and fill out an implementation report
... And then exiting CR is based on 3 things: 0 outstanding issues, have a test suite, and 2 or more implementations of everything in there
... THat goes for the other CRs as well
... This isn't anything special for as2
<aaronpk> someone is causing a horrible echo
tantek: Still a lot of work to do
... Where are we with th etest suite?
eprodrom: What we have for the test suite, two sides, one is the validator, which lets publishers check and see if what theyre publishing is valid as2
... The other half of that is the document suite that consumers can use to verify that they're able to consume legal activitystreams
... The consumer documents, you can run the consumer test suite through the validator and it all validates
... THe question is do we have for test coverage in the consumer documents for all the aspects of the spec
... that sounds like an excellent task for me to take on this week, to make sure that's the case
... that we don't have any features that are called out in the document that are not explicitly testable with documents
... I'm going to take that as an action for myself
tantek: when we reach the end of CR, we can evaluate the reports, if we have two or more implementations of every feature that's great, but if that's not the case we have multiple options at that point, including potentially dropping features
... Which definitely means we need to do another CR
... but that's a legitimate thing to say, we went to CR and nobody implemented it so we're going to drop it
... that's a perfectly fine thing to be able to say, keeps specs tight and lean
... It will merit discussion on the part of the working group, especially as our time dwindles, the chances of us exiting CR with this spec is greater if we drop features than wait for implementations, if it comes to that
eprodrom: do we have more to talk about with as2?
micropub i18n issues
aaronpk: Quick question about this process
... we know that we published as CR before addressing the issues from the i18n group
... Some of those are now incorporated into the latest ED, some still need to be done, but I'm wondering what we need to do about this, are we planning on publishing a new CR with these changes? What's the timeline on that? What should we plan on?
... This question is probably for sandro
... We published before addressing i18n issues with the goal of doing that later. Should we do that sooner?
sandro: were we expecting those changes to be normative?
aaronpk: going to describe things like how to do unicode in strings and things like that
sandro: if I'm remembering correctly the hope was that that wouldn't be normative
... as soon as possible we should figure out what the changes are.
tantek: I understand they are normative because they impact what implemenations SHOULD or MUST do
... have to treat strings in th efollowing way
sandro: we should do them as soon as possible. If the question is whether to wait ... if we know exactly which changes, do we wait for other possible changes too before reissuing CR?
... What's the deadline?
tantek: any other issues?
aaronpk: the only one that might be normative is the clarification about the reference to indieauth which chris opened
... the way it was worded in the CR draft was that it's a normative reference, which we should have caught because we can't actually normatively reference it
... I already updated the ED to say that implementations MAY use indieauth to get an access token, don't know if tha'ts normative though
... I think it said SHOULD before
tantek: and thatw asn't raised in transition call?
aaronpk: yeah somehow we just missed it
... Oh no I'm wrong, that change made it into CR draft
... Never mind
... But I am waiting for chris to thumbs up that
<Loqi> [Aaron Parecki] Micropub
tantek: in response to do we wait for more issues. Micropub has been CR for 3 weeks or so. Have you received any new issues from implementors that are normative?
aaronpk: there's a handful of issues, cweiske took a look at it and is implelmenting a server. they look like clarifications. I don't think they're normative changes
... I defintiely need to incorporate those changes, but I don't think tehy're normative
aaronpk: essentially yeah
... one example is when you're making a json request the spec should require the content type header
... you'd probably be doing that anyway but the spec iddn't say it
tantek: technically any changes like that are normative
... I'm not sure how that impacts CR exactly
... where it's a refinement, not actually a change. Narrowing the space of what implementations can do
sandro: the way ralph would interpret it is whether there's need for particular people that would affect. If it would affect implementers, are they okay iwth it
tantek: if there are implementations that used to be conformant that would no longer be
... can you take that as an action aaronpk? if any of those refinements would make existing implementations nonconformant?
... I think sandro is saying that if not that would not impact our timeline
aaronpk: Okay. I suspect theres going to be some that I can just incoroprate and close without worrying about existing implementations. I think some will affect implementations. My homework for this week will be to categorise thsee issues
tantek: that kind of refinement could be turned into a test. WHich would influence implementations
... And about AS2 as well, until we have a complete test suite for all the features, as you create tests you find that the spec wording is not entirely as you intended and so requires changes to its wording
... I know evan is already on it, encourage to evaluate the completeness of the test suite for as2. For micropub as well, the sooner you're able to produce a complete test suite the greater the confidence that we as a group are able to epxress that every feature has been expressed ina test so is testable so is written properly
... at least in CSS every time we've made a test suite we've found we've had to change the spec normatively
eprodrom: fair point
aaronpk: I noticed that as2 took the i18n and incoropriated the unicode rather than adding a new property. That made it through the draft and is in CR now
aaronpk: So I would like to essentially do the same thing with micropub to close issue 37
tantek: anyone object?
eprodrom: what you're saying aaronpk is that you want to do the bidi stuff in micropub the same as we did in as2?
<eprodrom> PROPOSED: add language to micropub to use bidi control characters and thus close issue #37
aaronpk: that works
<tantek> (consistent with AS2)
<tantek> (consistent with AS2 CR)
<tantek> might be good to add that specifically to the resolution to indicate the motivation / deliberate consistency in the group
sandro: why does html not work for this?
aaronpk: you'r enot always posting html in micropub
... could be plain text
<tantek> +1 with that explicit note of motivation of consistency with AS2 CR
sandro: that are defined in the spec as being plain text?
eprodrom: please vote. if you feel like you don't u nderstand at this point and want to keep discussing, that's okay
... don't feel rushed
RESOLUTION: add language to micropub to use bidi control characters and thus close issue #37
eprodrom: any more i18n issues on micropub we could discuss?
aaronpk: no that's it
eprodrom: anything more to discuss about micropub now?
<tantek> aaronpk, when I look at https://github.com/w3c/Micropub/issues I see plenty of open issues
<aaronpk> tantek, yes those are mostly feedback from cweiske that I mentioned earlier. I will be reviewing those and tagging any that need group feedback for next call
tantek: we should consider... if we are, and I think we should optimistically presume we have implementors implementing all of our CRs, we should have consistency with these kinds of details
... this change to micropub which brings it closer into alignment with AS2 it maybe worth issuing as a new CR for that reason
... so implementors reading them can see they're all designed to work in similar ways
... So even though there are other issues we know aaron is still processing, this might justify a new CR as soon as that edit is made
sandro: I don't know whether we need director approval to do a second CR, I don't know offhand
tantek: I don't think we do
... I think it extends the CR period by the minimum amount
<Zakim> tantek, you wanted to discuss CR consistency
tantek: I know it's a lot of work to produce a new version, but it may help implementors who are implementing multiple specs to get it right
eprodrom: aaronpk, what would that mean for you?
aaronpk: I do believe that finishing the test suite will probably lead to more changes in the spec, so I don't knwo if it's worth planning on publishing two new CRs or if I should wait until the test suite to publish a new CR
tantek: do you have a rough estimate4
... if it's like a week or two then we can wait, but if it's like a month then it might be worth doing a CR now or in a week when you can incoroprate that issue
sandro: doing another CR is hard
... because it requires a new call for exclusison to the AC and a new director approval
... Maybe that can be done via email
tantek: that sounds right
... It doens't require an actual telecon, but does require director approval by email, that sounds right
... When we explain what the difference is, that we're doing a new CR based on resovling horizongal review issues, the director should be very supportive in my opinion
sandro: and as you say it extends the deadline at least 4 weeks
tantek: which if we're not done with the test suite we will need to do anyway
... we're not going to be able to exit cr anyway until we're done with the test suite
... so if you think aaronpk you can finish the test suite quickly
aaronpk: I suspect ti will be after the week after tpac
tantek: that seems sensible
... you'll also get chance to meet with horizontal reviewers at tpac
... another good reason to have CRs as close to what they have asked for as possible
... You said it was hard sandro but I think we're able to handle that
tantek: and being responsive to horizontal review shows good faith behaviour on the hpart of the wg
sandro: I'm afraid we're running very low on a lot of resources
... I think we can focus on implementation reports from here
eprodrom: any editors have updates to give?
<tantek> up to you aaronpk, if you think you can produce a CR draft for next week with i18n issues resolve and want to, go ahead
cwebber: No update on spec from last time, but implementation stuff I'm pressing very hard on and have most of the client to server stuff in place, and a separate client being developed
eprodrom: Closing, thanks everyone
... back on regular schedule, we'll have a meeting next tuesday, tantek chairing
<ben_thatmustbeme> See you all then
<tantek> rhiaro++ for minuting!
<eprodrom> trackbot, end meeting