Social Web Working Group Teleconference
28 Feb 2017
See also: IRC log
- bengo, aaronpk, ben_thatmustbeme, cwebber, sandro, tantek, jasnell, csarven, eprodrom
- Ben Roberts
- Summary of Action Items
- Summary of Resolutions
<cwebber> I don't hear anyone yet
<tantek> great! thanks ben_thatmustbeme
<scribe> scribe:Ben Roberts
reconfirm next telcon
tantek: we are currently schedules for 3/14 are there any objections to that?
sandro: one thing we should mention is the meeting time, i don't know if you have noticed KevinMarks complaining about it
tantek: i did
sandro: though I don't take that as seriously since he wasn't on every meeting in the other time slot
tantek: he made some, but this time is better for Amy and also probably better for Europeans
... if Kevin raises a serious objection we'll deal with that
... if anyone gets the sense its more serious, we will deal with it
... lets say that the march telcon time is confirmed
approval of past telcon minutes
PROPOSED: approve minutes of 2-14
<bengo> +1 to approving those minutes
RESOLUTION: approve minutes of 2-14
<csarven> Oh, just realised there is a meeting.
(discussion of order of meeting agenda)
<cwebber> I +1'ed twice
<eprodrom> ben_thatmustbeme: thanks!
<cwebber> AP can be short
Micropub CR to PR
tantek: i put a meta-item for it to the agenda for any CR to PR discussion
... the questions are 1) test-suite ETA and imp. coverage, 2) ... (refer to agenda)
aaronpk: unfortunately the test-suite has not made any new progress. It is only a client test, and i realized taking time to do that would push the entire schedule back. So instead i created a implementation report template
... as of this point there are 11 client implementation reports are submitted
aaronpk: i put together a summary just like webmention as well
... dark green is more than half of the implementations have implmented it, light green is at least 2, and anything with only 1 is yellow
the only thing that has only 1 is a vocabulary that was just put in for curiousity
aaronpk: (the link) is the server report as well. I should probably do it as a spreadsheet as well
... for example 20, 21, and 22 are all one for example, in the spreadsheet version those would all be in one row
tantek: so those are all ways you can do it? i'm trying to understand
aaronpk: yeah, they are different ways to recognize a successful update
tantek: does that mean that clients must handle all 3
aaronpk: exactly, in the client report there is a line about it handling all 3
tantek: that makes sense to me, if its looser on the server side its stricter on the client side
... it would be great to see the spreadsheet version of this as well
tantek: it looks like there are at least 2 implementations of each feature
aaronpk: yes, we have had that for a while
sandro: i'm on the client report, its great that you made all those implementations but i dont' think you having 2 implementations of a feature should really count
... just looking down the rows there is only 1 feature that looks to have that issue
ben_thatmustbeme: i have plans to update one of my clients which is out of date for that
aaronpk: there are a few that i could ask to go and implement it
tantek: i think it would be arguably challenging to have two implementations from the same person. you might be making the same decisions even if they are the same code base
sandro: it would be nice to not count your own, the editor has a short-cut which is their brain
tantek: sandro, are you proposing we don't count editor implementations at all?
sandro: i'm not, its raising the bar late in the game, but it would make a stronger case
tantek: does that mean we should delay going to PR?
sandro: i'm not really proposing that
... this sort of hits me as odd but its because so many parts are only optional
<cwebber> also note that we're getting to the halfway mark on the hour and we have a *big* topic today with AS2
aaronpk: yes thats because its possible and very useful to have clients that only support creating, if the server supports that at all
<aaronpk> cwebber, but we didn't really started until about 15 minutes into the hour anyway ;-)
tantek: this reminds me a lot of the as2 report where different implementations use different sets
sandro: it kind of surprising to me in a protocol to have that though, it sort of gives me the sort of thin-ice feeling
... it doesn't mean there is anything wrong with it, it just doesn't give me the security
aaronpk: i could also group the categories by those who actually implement updating at all
sandro: i guess the question is, do you feel like those 2 implementations from outside the group are good enough?
aaronpk: i do
sandro: thats sort of on the WG to decide if thats enough
tantek: to that point i feel like external implementations carry more weight
... when i see multiple implementations from outside the working group, i feel good about it, when i see only one, i am concerned, when there are none, i am very concerned. not that i would stop it, but I would find it very concerning
sandro: i feel like this group is on more thin-ice than i am used to as usually everyone is paying attention to all the specs and this group is more clustered to people only looking at some specs
tantek: so whats the summary of the implmentation status given all of this, i suppose there is the one sandro pointed out
aaronpk: that one would be certainly good to get another implementation that is not me
cwebber: i am just going to suggest that we set a time limit on this so that we can get to AS2 since that is why we scheduled this meeting
aaronpk: i am okay with that, but micropub did get bumped
sandro: we also agreed this was a 2 hour meeting
tantek: i am getitng the feeling we are not ready for CR based on this one feature?
sandro: i agree but i don't think we need a meeting to do that, maybe we could approve pending that
... that said i wouldn't stop it based on that, but it makes a better case when going in to the meeting on it
tantek: so it sounds like it may be worth waiting
sandro: i think its worth waiting a week or two
tantek: so 2+ implmentations one of which should be not the editor's
... so we could agree that we propose with it pending that one item
aaronpk: i would prefer to go with that
tantek: any issues needing group discussion?
aaronpk: no, changes are documented in changelog on editors draft
... they are editorial clarifications
tantek: sounds like you have ticked all the boxes, anyone else?
eprodrom: i wanted to ask, with that proposal, if there are any other changes in the time we are waiting for that implementation....
aaronpk: you mean changes in the spec?
tantek: its possible at any time that someone can file a new issue, we can't control that
PROPOSED: move micropub to CR pending an implementation of query for a single property
<sandro> that should be PR not CR
<sandro> PROPOSED: Move MicroPub to PR, pending an implementation of query for a single property (from someone other than Aaron)
RESOLUTION: Move Micropub to PR, pending an implementation of query for a single property (from someone other than Aaron)
post type discovery
tantek: i hope this will be fast as its just a request to publish an updated WD
<eprodrom> chair: eprodrom
tantek: it resolves some issues on github, it has a few minor fixes
eprodrom: your description is pretty straight forward, are there any questions from the group?
<eprodrom> PROPOSED: publish a new working draft of Post Type Discovery based on current editor's draft
eprodrom: if there are no questions from the group... i think we can move to the proposal
<aaronpk> +1 no questions, looks good
RESOLUTION: publish a new working draft of Post Type Discovery based on current editor's draft
eprodrom: are there other points on PTD you'd like to bring up during the meeting?
tantek: i think there were a couple issues i was waiting for commenter, let me pull those up to see if there are any that are worth the groups time
... one of the issues that i resolved with consensus in the thread was issue 13, which is waiting for response from the person to say that its ok
tantek: i just want to get confirmation from the group that this is a good resolution to this issue
eprodrom: so its been waiting for commenter for a while and now we are looking to close it
<eprodrom> PROPOSED: close issue #13 of Post Type Discovery as resolved
tantek: yes, per our github workflow we wait for original commentor to close it, or we get a group proposal to close it
<eprodrom> PROPOSED: close issue #13 of Post Type Discovery as resolved since text was added to the document
RESOLUTION: close issue #13 of Post Type Discovery as resolved since text was added to the document
<eprodrom> chair: tantek
ActivityPub to PR
tantek: as we did with micropub can we go through the list of bullet points
cwebber: i've been pushing hard to get implementations, i have a large set of features implmented and groundwork for the test suite, i have been working on mostly implementing to help AS2
bengo: i just want to ask this as an activitypub implmentor, is there any real chance that this will be a REC? I don't mean to offend anyone, but we have such a limited amount of time, is it better for the CG?
tantek: the WG believes the spec is ready for implmentations, we are trying to make sure that the spec as written is implementable, there has been a bunch of work there
bengo: i think cwebber and I are the only ones that have started implementing, and there isn't even an report template yet
tantek: cwebber would you be able to make up an implementation report in the next 2 weeks?
??: its in the PR
<csarven> I think rhiaro has (part?) AP implementation
<Loqi_> Abasset made 1 edit to Socialwg/2017-02-28 https://www.w3.org/wiki/index.php?diff=101642&oldid=101559
<Loqi_> Sandro made 1 edit to Socialwg/2017-02-28 https://www.w3.org/wiki/index.php?diff=101643&oldid=101642
<Loqi_> Tantekelik made 1 edit to Socialwg/2017-02-28 https://www.w3.org/wiki/index.php?diff=101644&oldid=101643
<Loqi_> Aaronpk made 1 edit to Socialwg/2017-02-28 https://www.w3.org/wiki/index.php?diff=101645&oldid=101644
<bengo> I understand and respect your time constraints cwebber, just forcing the issue
<bengo> Happy to move on w/ agenda now that I've poked a bit. cwebber++ for work so far.
cwebber: sorry, thats my fault, I will look at what bengo did and i will build on that. I also will say that the implementor behind mastadon is planning to implement, the intent is to get it done
tantek: certainly, we appreciate all the work you have done, the intent is to get it to REC, but the focus has been more on AS2, i expect that AP will see increased activity in the next month
cwebber: i appreciate that bengo is trying to push things forward too, i should have a bunch more information by next time
tantek: we have an agenda for next telcon, i'll leave it to you cwebber to add it to the agenda there
tantek: i'll let you take it eprodrom
eprodrom: is amy on the call?
eprodrom: 2 weeks ago the concern we had was that we were going to miss some of the features for AP and annotications because we didn't have implementations for them
<sandro> a bit out of date: https://www.w3.org/2017/02/social/implementations/as2/
eprodrom: and we had them at risk really. the good news is that we have had a number of new implementation reports come in over the last few weeks, several in the last 24 hours
... i was trying to run amy's script but have been getting an error, i think we are pretty confident that it is no longer a concern
<bengo> I will scribe
<bengo> how do I scribenick
<eprodrom> ben_thatmustbeme: there's a bug that's throwing an error
<scribe> scribenick: bengo
eprodrom: If your python is strong cwebber you may be able to fix what I couldn't
... Having a new version would help with discussion
<sandro> contentMap, etc
eprodrom: Second concern we had that didn't come up last week but did come up since is the feature of languageMaps. Feature where instead of having simple strings for some values, there's an object with language code to -> string mapping
<sandro> cwebber, it's linked from https://www.w3.org/2017/02/social/implementations/as2/
eprodrom: We only had one implementation of that as Publisher and Consumer
... I did one implementation of this for node.js impl. It should cover these features
... They were high priority because they are the i18n mechanism
I can implement this for distbin.com if it will move the needle.
<tantek> wow that's great!
eprodrom: Fortunately we do have implementations for these.
... The 3 lang maps to have 2 impls both publishers and consumers
... We've come to the last sticky wicket.
eprodrom: We've discussed several times the criteria for existing CR. When we discussed last week, we remembered that in previous meeting we talked about requiring 2 publishers and 2 consumers for each feature of the spec.
... We had not actually raised that to the level of a proposal or resolution.
... I think we had talked about it informally, but it had not come out to modify the exit criteria
... We had expected to have this requirement, but the exit criteria says differently
... It would be nice if this requirement was not a material difference.
... On the implementation report now. With the implementaitons that came in this week, I believe we're fully covered for the ones that are light green. Which is good news.
The implementation report doesn't ask, for each feature, whether you are a publisher or consumer
just if you are *overall*
and then it adds 'PC' to every feature you mark as 'y'
eprodrom: It comes down to what we want to do as a group
... My inclination is that we should make a change to the exit criteria to make it 2 pubishers and 2 consumers. Then we don't need to push it anymore.
... Then we move as expected, any features that dont meet that
... It would let us move on
I would like to stop scribing and q+ to discuss my previous comment
tantek: sandro what do you think of eprodrom proposal?
sandro: I'm torn. I like having more implementaiton reports. But I don't see any evidence that we agreed on this. I couldn't find anything in minutes, and I looked for awhile. It's been like a year so I don't remember our discussions. But I can see the minutes.
sandro: I've been in lots of WGs that don't use that standard. So I think I would have noticed
sandro: My inclination is to keep it at the lower bar
sandro: That's clearly good enough in other places
tantek: Can you cite a place
sandro: Web Annotations
sandro: I think so. Their bar is that consumers are enough.
tantek: I read their exit criteria. And despite their saying nothing about doing something meaninful. It did mention you must consume it and produce valid triples, etc
tantek: From my reading it sounded like they wanted 2 producers and consumers
<ben_thatmustbeme> eprodrom, https://ben.thatmustbe.me/static/reports.html
<Loqi_> ben_thatmustbeme has 61 karma in this channel (184 overall)
<eprodrom> THANK YOU
bengo: i wanted to point out that as evan said, looking at the reports as they are rendered now, don't give us an accurate reflection of that
<cwebber> Pubstrate does implement Mention btw
<cwebber> I must have missed it :\
<cwebber> on the IR
<tantek> cwebber, good thing we have the colors highlighting this! :)
<cwebber> also some of these items weren't on the IR page, esp the AP specific ones
bengo: it doesn't ask you on the report if you are publishing and consuming on each, its just asking if you once
<cwebber> eg inbox, endpoints, etc, were not on the template
<cwebber> they aren't in AS2 core tho
<cwebber> those are AP extensions
eprodrom: do you know of an implementation that publishes some features and consumes others
bengo: I'm sure mine does, but i think if we want to change exit criteria it should take that in to account
<sandro> I see only 'inbox' and 'outbox' are light green. Used by sloph (Amy) and distbin.com
<bengo> it me :)
bengo: i don't want to change exit criteria with bad data in front of it
<cwebber> can I reply to that
<cwebber> I don't think they are in AS2's terms
sandro: it looks like inbox and outbox are the only things, and since thats your implementation, would you be okay with that being an extension?
bengo: i would be, but if we change the ..
<bengo> FWIW I said I would be okay with those not being in AS2
cwebber: it makes sense that those aren't there since they aren't actually in the implementation report, and they also aren't even part of AS2, they are part of activity pub!
<sandro> +1 add "FYI" note to extensions and stuff like that
tantek: i think i mostly agree with cwebber, with the other implementation reports summary table, we've put the extension type things at the bottom so its more just additional FYI, its certainly not required, but i think its good signaling, of "look at how green the spec is, and huh here's some interesting extensions in there i should look at"
... if its possible, could we do that? and i'm hoping whatever generates this could do that as well
cwebber: that sounds good to me, the bonus implementation report items
eprodrom: so i don't want to have anyone file implementation reports again, it would be great if we could change the report creation script so you could specify i'm only publishing or consuimg this feature. I don't think we are really going to have a feature thats going to come up where thats material
... some of these are libraries so they are just dealing with these in a similar way, so i think its unlikely that they will be dealing with them in a seperate way from consuming and publishing
<cwebber> +1, would love to have it, doesn't need to hold up implementation
eprodrom: looking throught the report now i'm not seeing any case where we have 2 publisher and 2 consumers, but i don't think thats the case
eprodrom: i don't want to hold up publication for that
<cwebber> not Mention
tantek: let me summarize that, are you saying based on the report you are determined of what makes sense to drop to go to PR?
<cwebber> I just submitted a PR
<cwebber> it wasn't in the template, but Pubstrate implements Mention
<sandro> Are we all looking at https://ben.thatmustbe.me/static/reports.html
eprodrom: yes, all of those that are in red and probably some in light green as they are not part of the spec
<cwebber> I can reply
tantek: i only see one yellow which is 'mention', can you clarify that?
eprodrom: i wouldn't wait for more implementation reports on it, honestly
cwebber: that was another one of the things that was missing from the template and just got added.
... it is implmented by one more thing than is listed there
<bengo> There will still be only 1 implementation
<bengo> there are 0 now right?
<bengo> my bad
eprodrom: that one would go green at that point
<sandro> I just copied https://ben.thatmustbe.me/static/reports.html to https://www.w3.org/2017/02/social/implementations/as2/
<bengo> sandro thanks
<cwebber> I don't use it in any meaningful way
eprodrom: i think the only main item in red that i would be concerned about would be ... relationship. there wasn't a place for it in the implementation report
<cwebber> I think that's just: IsScontact, IsFollowedBy, IsFollowing, IsMember
i haven't seen it in ay of the implementations that came in
<bengo> got it
<cwebber> er IsContact
<bengo> ... i dont use and never have
sandro: is it also followers and following?
<cwebber> I'm okay with that being axed personally
<cwebber> it could be an extension
eprodrom: thats from AcitivtyPub
<cwebber> if someone needs it
<tantek> I'm seeing the following in red: IsContact, IsFollowedBy, IsFollowing, IsMember, as, authorizeClientKey, endpoints, followers, following, oauthClientAuthorize, preferredusername, provideClientKey, proxyUrl, source, uploadMedia
<cwebber> tantek, and everything that isn't the Is* things is part of activitypub
eprodrom: the proposal that was on the table was to make a change to the exit criteria, i'm not sure what happens when we change that before going to PR
<tantek> thanks cwebber, appreciated.
<cwebber> I don't see any reason to hold off on going to PR
sandro: we can't change our exit criteria, we can remove things from the spec and hold off going to PR as a group, we can do that just among ourselves, similar to what we did with micropub
<bengo> Can we... remove the red properties, then propose to raise the criteria bar given that all the remaining properties will still be in the PR, then go to PR?
<cwebber> the AP items don't show up on the previous report
tantek: there are 2 things here, 1 is to agree to go to PR and to get these in to a state that would allow us to get this in an clearer form
<cwebber> I'm not sure why they show up on the new, generated report
tantek: not all of this work has to be done before we vote to go to PR but before we ask our staff contact to take it to w3c management
... normatively speaking, we have to drop the relationship vocabulary stuff ... but let me go to queue
cwebber: the last generated as2 report didn't show inbox, i would suggest we just remove those or move them down
<cwebber> +1 on that
eprodrom: i have opened up an issue on the report generator to filter out properties that are not part of as2 in to an extensions area
sandro: can we just have a resolution to remove those 4 items we specified?
<tantek> I'm seeing a bunch of yellow here https://www.w3.org/2017/02/social/implementations/as2/ that I'm not seeing on https://ben.thatmustbe.me/static/reports.html
<bengo> and Mention?
<sandro> bengo, cwebber said he implemented Mention but didnt report it because it wasnt on template
<bengo> sorry thanks tantek
PROPOSED: drop ask-risk terms, isContact, isFollowedBy, isFollowing, isMember from AS2
<sandro> +1 since no one has reported implementing them, it seems safe to drop
RESOLUTION: drop ask-risk terms, isContact, isFollowedBy, isFollowing, isMember from AS2
tantek: now that we have resolved to drop those as normative terms, do you want to consider them as extensions? are these still a good idea? we put them in the CR, but no one implmented them
eprodrom: the reason we had these was that we had some issues in the group that we had already had ... (cracked up a little) ... referring to external vocabulary, we should have a simple way to bring it in
<sandro> The report shows nothing from activity-streams.js which suggests the report is not being processed properly
eprodrom: it appears to be something that we are not putitng a lot of implementation in to, adding those as another external vocabulary, would not be very helpful
sandro: i think we can vote on go to PR then
tantek: if you are accepting what evan is proposing
sandro: i don't understand
tantek: eprodrom point was that we do have multiple implementations for 2 producers and consumers
sandro: i don't see a need for that, why would we even talk about that
tantek: i agree with eprodrom that we update it that we had that was our original intent
sandro: no, its not minuted there
<Loqi_> [sandro] RESOLVED: Our CR-Exit for AS2 will be: Each feature of AS2 must be used by at least two independent implementations
sandro: why do you want to spend time on this issue, what does it accomplish?
tantek: it strengthens our spec and continues a good practice of ..
sandro: the exit criteria will be dropped in the PR draft
<csarven> I'm actually hearing that everything is fine
<tantek> the "used" in the minutes is the part I'm seeing as requiring 2+ consumers
eprodrom: i think we its a moot point with the point of what features we roll out
<eprodrom> PROPOSED: recommend going to PR with all features that have 2 publishers and 2 consumers
<cwebber> it doesn't say 2 publishers 2 consumers in the thing that was RESOLVED
<cwebber> it just says 2 implementations
<cwebber> linked from minutes bengo posted
eprodrom: what i would propose is that we not change the exit criteria but we do this in our final vote
<tantek> it says *used by*
<bengo> -1 we should go to PR with what the exit criteria says...
<bengo> because that's what they're for?
<cwebber> suggested language!
<cwebber> I'll type it out
<bengo> go to PR with all features that haven't been removed?
eprodrom: i'm not sure i understand what standing on principle will help us here, if we can find some language that appeases everyone here, we are just splitting hairs here
<cwebber> yeah that's good
<csarven> +1 to Sandro's version
sandro: but i dont' even know how to parse that, how about we just say we vote to go to PR
i don't know what that means
<cwebber> how about PROPOSED: ActivityPub move to CR with edits made in this meeting (dropping Is* terms) on the basis that all are satisfied that we have two implementations used of each term.
sandro: does that mean that we are deciding to publish or not
<bengo> cwebber is good
<cwebber> PROPOSED: ActivityStreams to CR with edits made in this meeting (dropping Is* terms) on the basis that all are satisfied that we have two implementations used of each term.
<cwebber> PROPOSED: ActivityStreams to PR with edits made in this meeting (dropping Is* terms) on the basis that all are satisfied that we have two implementations used of each term.
<sandro> +1 !!
<tantek> PROPOSED: Take AS2 to PR with all features that have 2 publishers and 2 consumers, dropping terms that don't meet that as noted in https://github.com/w3c/activitystreams/issues/406
<sandro> -1 tantek
<eprodrom> +1 and +1
<bengo> that's not the same tantek
<eprodrom> Tastes great AND less filling
tantek: i'm trying to capture evan's proposal
<eprodrom> Drop my proposal please
sandro: don't, let evan capture evan's proposal
<eprodrom> c uip;
<bengo> eprodrom dropped his proposal. The next proposal state is cwebber 's
<tantek> PROPOSED: Take ActivityStreams to PR with edits made in this meeting (dropping Is* terms per https://github.com/w3c/activitystreams/issues/406) on the basis that each term is used by 2+ implementations.
<csarven> stop. break dance.
RESOLUTION: Take ActivityStreams to PR with edits made in this meeting (dropping Is* terms per https://github.com/w3c/activitystreams/issues/406) on the basis that each term is used by 2+ implementations.
<cwebber> lol :) yay
tantek: 2 resolutions to go to PR in one call
<cwebber> just drop 'em
sandro: does anyone want those 4 dropped terms to be an extension or completely drop them?
tantek: yes as i understood it, we are just dropping them completely
tantek: just an FYI, annotations kept their exit criteria as a note
<cwebber> just a comment
cwebber: i suggest we remove the relationship example in there with an external URL
sandro: do we have an example of someone who is actually using that
eprodrom: we were originally using relationship from (link), my intention is to just replace it with one of these
... i'll put a note on the issue
tantek: this is an informative example correct?
ben_thatmustbeme: i'll do my best to update the report generator if we can get some updates to those messed up reports and someone can send me a list of extension items from this list
sandro: do we want to drop the empty column?
tantek: lets move that to the issue that was opening on improvements to the report
<cwebber> ben_thatmustbeme, here are the AP terms to be marked as extensions: as, authorizeClientKey, endpoints, followers, following, inbox, oauthClientAuthorize, outbox, preferredUsername, provideClientKey, proxyUrl, source, streams, uploadMedia
<eprodrom> That's what I was asking for
sandro: one minute on the F2F, there are only 2 people that have said yes to any specific date, that does not bode well
<aaronpk> FYI I just published the micropub server report summary https://micropub.net/implementation-reports/servers/
<cwebber> sorry, hadn't replied earlier
<eprodrom> I just added my times; thanks for sharing it
(discussion of likelyhood for f2f)
tantek: we can try again for may, that would be the last possible chance for us though
... what do people think about trying to do an F2F in may?
sandro: would it be any different for may?
tantek: i don't see how we are going to come to a conclusion on this, other than it seems unlikely we are going to do this in april
<sandro> cwebber, and it turned out to be super long anyway
<Loqi_> tantek has 49 karma in this channel (320 overall)
<eprodrom> thanks so much for handling this great discussion
trackbot end meeting
Summary of Action Items
Summary of Resolutions
- approve minutes of 2-14
- Move MicroPub to PR, pending an implementation of query for a single property (from someone other than Aaron)
- publish a new working draft of Post Type Discovery based on current editor's draft
- close issue #13 of Post Type Discovery as resolved since text was added to the document
- drop ask-risk terms, isContact, isFollowedBy, isFollowing, isMember from AS2
- Take ActivityStreams to PR with edits made in this meeting (dropping Is* terms per https://github.com/w3c/activitystreams/issues/406) on the basis that each term is used by 2+ implementations.