Social Web Working Group Teleconference

25 Oct 2016

Attendees

Present: eprodrom, aaronpk, wilkie, akuckartz, ben_thatmustbeme, rhiaro, sandro, cwebber, tantek, bengo
Regrets
Chair: Evan
Scribe: ben_thatmustbeme

eprodrom: we do have our F2F coming up in november, 3 weeks away, if you have not yet signed up please do so, so we can plan logistics
... link in IRC
... it will likely be one of our last F2F meetings, if not our last and we have a lot of talk about there
... sando, do we have anything else to discuss there?

sandro: not unless anyone has questions

AS2

eprodrom: we had a lot of discussion of AS2 going to CR last week
... we have a few bugs to fix, and some other items

<cwebber2> eprodrom, I should be able to get implementation reports in, though I might need to be pinged

eprodrom: i did not have much time to work on it this week but i will be pushing it this week
... mostly its editorial issues, a few normative issues that we resolved that i just need to close, and we need implementation reports
... i would really like to have our first implmentation of pump.io by F2F
... i'll also be doing an implementation report for the validator and hopefully james will have one too
... there was a question about twitter using AS2. i reached out last week, but I am still waiting for a response back
... are there any questions for me or james on AS2?

<rhiaro> I just want to raise for myself mostly that a human readable version of the AS2 namespace is still on my TODO list

eprodrom: if not i'm good moving along for now

<rhiaro> (and adding the extensions section with AP terms, etc)

eprodrom: <<repeats rhiaro comments from chat>>

ActivityPub

<cwebber2> shoot

<cwebber2> my phone disconnected

eprodrom: lets hold off on that while cwebber2 reconnects

PubSub

aaronpk: I did apparently volunteer last week to create the test suite, so far i have a plan of the list of tests to write but i plan to make progress on that this week.

<cwebber2> reconnected

aaronpk: over the last 24 hours or so there have been a whole bunch of issues opened on pubsub
... i would love to get more eyes on them

<sandro> https://github.com/w3c/pubsub/issues

aaronpk: i tackled half of them already this morning to try and get some resolved if possible
... link in IRC

eprodrom: any in here you find needs discussion today? or are we not ready for that yet?

<Loqi> [@jirikuncar] Welcome PubSub protocol standard on W3C 🎉 https://www.w3.org/TR/pubsub/ (http://twtr.io/1JVB3kx39eC)

aaronpk: there were some suggestions to rename it... again

sandro: lets leave that to the thread

<tantek> oh dear really?!?

aaronpk: i think most of these are fine to allow to continue to talk on the threads

<wilkie> if that's the biggest issue that doesn't seem so bad haha

<Loqi> awesome

<aaronpk> new name issue: https://github.com/w3c/pubsub/issues/10

eprodrom: i really hope we don't persue that much further, the name took enough effort

<tantek> FWIW I can live with "PubSub" and have yet to hear anything even remotely more palatable

eprodrom: on the test suite, do you have an intention to do an online test server similar to the web webmention works?

aaronpk: yes, similar to webmention and micropub, a tool that acts as the other end for you

and will track what features are available and build the implementation report from that

eprodrom: have you had a chance to look, are there normative issues here?

aaronpk: there are some around timing and retrying
... some are editorial, some are tightening up language
... like defining that a time in seconds must be positive integers, things like that

eprodrom: so some are normative, okay
... anything else on this agenda topic?

sandro: can i ask one big question?
... when we first talked about it, i raised the issue of payload type. have we thought more about what we are doing there?

<aaronpk> https://github.com/w3c/pubsub/issues/24

sandro: i haven't seen any discussion, but maybe it happened...

aaronpk: no, we haven't. There is one issue around diffing versions. I don't think its going to be possible to actually specify that around that
... i think this is going to be around notifiying and not around the actual content delivery

sandro: so there is going to need to be another spec?

aaronpk: i think there will need to be a profile of the spec, similar to oauth2

sandro: so for me to use this to syndicate blog posts, this isn't going to tell me how to do it?

<aaronpk> https://www.w3.org/TR/pubsub/#publishing

<Loqi> [Julien Genestoux] PubSub

aaronpk: you will need this spec and the documentation from the hub you are using

sandro: that seems sub-optimal

<tantek> does anyone actually use the payload in practice?

aaronpk: i agree, also the publishing section is empty, which is a seperate issue, but similar

sandro: that leaves me confused as to what the use of this spec is

aaronpk: it provides interop on some aspects

tantek: the flip side is that we have real world experience where it is working

sandro: but thats not via this spec

tantek: i don't think you are asking the right question

sandro: i think there is clearly something that works, but what works is not being specified by this spec

tantek: i'm hoping that we can fill that in based on experience

aaronpk: one of the issues there is that the spec tries to be content type neutral
... to do that, it can't specify that with diffing types
... the most interoperable way to do that, is to use XML feeds, specifically RSS and ATOM
... one way is to put those in to the spec
... but then its not really content type neutral

sandro: one way is to put what you just said in to the spec and then give people links to where they can handle those

tantek: i feel a little uncomfortable with that, i feel like some of it will have to be in the spec
... i think its a reasonable approach to document "heres the content-types that we have found to work as of today" and then give people some idea of how to make their own types work
... i guess from an implementor perspective, i'd rather document what works now, rather than dance around it for now
... as for right now i believe we have interop with RSS, ATOM, and h-feed

aaronpk: yes

tantek: and then it becomes a list for people to get into the party with

sandro: i guess it becomes a question of where we put that? normative, as another note?
... i'm guessing that h-feed is maintained in the same way as mf2?

tantek: yes, the individual format specs (RSS, Atom, h-feed) already document how to parse the formats, what we can do is specify what to do with that parse result, which tends to be very similar in each of those cases at least

<tantek> no I said the format specs document the parsing of the formats

sandro: I'm trying to picture how that text is, for RSS i feel that would be a paragraph

<aaronpk> https://documentation.superfeedr.com/subscribers.html

<Loqi> Subscribers

tantek, can you document that better

you spoke fast

aaronpk: let me link to the documentation for superfeedr, (in irc), let me say that content type is also only important for fat-pings
... with thin pings there is no content in the notification, then its outside of the spec for the subscriber to actually fetch the original URL, and then its all outside of this mechanism

eprodrom: it sounds like we could get a lot of use out of getting like 4 content types in there and that would get the 80/20
... it sounds like sandro needs to make an issue in the issue tracker

tantek: one of the nice effects of this is that we can treat each of these as somewhat orthoganally, and when we find one we can make tests based on that content type
... and we can just put any others at-risk
... i'll just be direct, i don't think anyone is publishing as2 with pubsub, we could put it at-risk and if we need to get through CR and we don't see any implmentations, we can also spin it off in to a note and then that note can be maintained by the community group or whatever

eprodrom: its been a while, but its things like "do you send the entry or the feed with one item in it" literally that can be solved with one sentence, and that would help a lot with interop

tantek: the only part i had on that was the different was the diffing vs the fat-ping

<eprodrom> No more pings!

tantek: the thin-ping is to send the URL only and then the diffing is seperate, if we use the fat-ping method, we have to specify what goes in to that body
... we would have to consider those two approaches independently as well

eprodrom: anything more to discuss on pubsub?

tantek: whats the next step?

eprodrom: sounds like the next step is to address the issue queue, i count around 17, with considerable discussion
... sounds like we'll be having some next steps

ActivityPub

<cwebber2> https://github.com/w3c/activitypub/issues/

<cwebber2> https://github.com/w3c/activitypub/issues/162

cwebber2: hello, last week we talked about bringing AP to CR this week. We had some really involved issues that we ended up having to go through. They are all now marked as 'waiting for commentor', the 4 that were filed by bengo. I don't know what the process is on this, i have answered them with i htink an acceptable response, with the exceptions of 162. I am not quite sure what was being asked

<tantek> hmm, we should get bengo here to discuss

cwebber2: we are waiting on those 4 issues, hopefully bengo will respond within the next week
... there are some other issues, but they seem like they will be easier
... there are 2 issues i feel should be discussed in the group

<tantek> !tell bengo can you join us in the telcon 2016-11-01? we need your help resolving some ActivityPub issues interactively so we can get move it along toward PR!

<Loqi> Ok, I'll tell them that when I see them next

eprodrom: so you have 2 issues that you need to be addressed by the group, lets try that now
... lets get those 2 first we'll then go on to SWP and we'll come back and see what time we have for additional ones

<cwebber2> https://github.com/w3c/activitypub/issues/155

cwebber2: 155 is potentially thorny. Someone has asked to have a history to show old revisions of an item
... i don't think that would be trivial method unless we do it in a history entry
... i feel its a bit late to add to AP core
... i think it makes sense as an extension

eprodrom: i think it makes more sense to have a collection of activities like lifecycle, here is its creation, its edits, comments, here's its delete
... so using the activities to show the history
... that said, if you wanted to have a history property, then it would make sense that way

<tantek> has anyone implemented this a all?

cwebber2: i would be okay to add that with marking it at-risk, but i feel like its likely not something thought through well enough
... do you think we should put it as "for later version" and leave it as an extension for now

eprodrom: i htink that makes sense

<Zakim> tantek, you wanted to note stage of an idea / proposal vs stage of a spec

<tantek> yes, for "later version" would be appropriate

cwebber2: is it okay to mark it as later version and close it or leave it open

eprodrom: i would leave it open

tantek: the general concern here is that activity pub is basically like trying to be in CR, even though we aren't. That means we really don't get to add new features unless its a major blocker for implementation
... once we recognize that, what level of maturity is it? is it an idea? prototyped? already working for people? technically we can add anything to the spec thats within the scope of the charter, but generally we have been trying to push for people to incubate them before putting them in a working draft
... from my understanding no one has even prototyped this

cwebber2: thats my understanding too
... great, i will respond to the issue as such

tantek: and encourage them and even possibly help with incubation

cwebber2: i'm reasonably optimistic that we will be able to go to CR next week

<akuckartz> cwebber2++

<Loqi> cwebber2 has 76 karma

SocialWebProtocols

<eprodrom> rhiaro: ?

<tantek> cwebber2++ thanks for your diligence with the flood of ActivityPub issues. My goodness.

<Loqi> cwebber2 has 77 karma

rhiaro: there was a request for me to update SWP with new doc status, we haven't had the transition calls for LDN or PubSub so i didn't know if we wanted to wait

tantek: i think its worth it for FPWD of PubSub
... we don't actually have the transition called scheduled yet do we?

sandro: we're working on it
... we may not need it

tantek: there are two things we need on that call

rhiaro: we are aiming to get them published next tuesday one way or another

eprodrom: are we ready to go to another WD with an update of FPWD of pubsub?

<cwebber2> tantek, thank you :)

eprodrom: if we do a resolution, you would be able to publish it

<rhiaro> http://w3c-social.github.io/social-web-protocols/#changes-since-2016-10-12

eprodrom: is there any reason not to do that, is there anything else we want in there?

<rhiaro> (that fragment is broken, willf ix)

<eprodrom> PROPOSED: publish a new working draft of Social Web Protocols

<cwebber2> +1

+1

<rhiaro> +1

<eprodrom> +1

eprodrom: amy is that ok?

<tantek> +1

<wilkie> +1

<rhiaro> There's a bunch of stuff that needs to go in there still, but if we're not waiting for other document status updates, there's no need to wati for that

<akuckartz> +1

RESOLUTION: publish a new working draft of Social Web Protocols

<tantek> I like this rhythm of updating our overview document when we update the individual documents.

eprodrom: you have the resolution you need

<tantek> I think it helps the outside confidence in our overview document, which is pretty important for understanding our work as a whole.

eprodrom: we are near the end of our agenda, i believe we have webmention next

<aaronpk> sandro, let me know if this captures it for you https://github.com/w3c/pubsub/issues/27

Webmention

<aaronpk> https://github.com/w3c/webmention/issues/67

aaronpk: the only thing i wanted to talk about today, we that in webmention we need to change a reference to the WHATWG spec instead
... its a more up to date spec of what a URL is, especially for hte purposes of webmention
... its technically a normative change, but it won't effect any implmeentations

tantek: its my understanding that this is the kind of change we can make between CR and PR

sandro: its a change in reference between what and what?

aaronpk: from RFC3986 to the WHATWG URL spec

sandro: whats the difference between them? i'd have to run it by ralph
... if you want to just send me a draft and CC chairs

tantek: i've seen the webapp grounps switch as well, so i think this makes sense, we're not the first group or spec doing this
... we can provide citations for that

sandro: and there is no intended difference?

<cwebber2> lol

tantek: it should not effect implmentations, but its a more up to date reference

eprodrom: is there enough of a difference here that its worth this effort?
... can we just gloss over this and move on?

tantek: i think its a due dilligence kind of detail, we're getting close to PR and this is crossing your T's and dotting your I's
... the intent was to cite what a URL was, not to specifically reference an RFC
... i think thats a good detail to fix before we go to PR
... its good that we're only talking about this kind of detail at this point

eprodrom: it makes sense to me, i just don't want to spend a lot of cycles on this unless its a worthwhile change, and it sounds like this is something everyone will be doing anyway
... do we have other discussions on webmention?

tantek: do we want a group resolution on that issue pending ralphs approval?
... we might as well get it out of the way so we don't have to talk about it again

<eprodrom> PROPOSED: adopt https://url.spec.whatwg.org/ as URL standard reference for Webmention

<tantek> +1

<eprodrom> +1

<cwebber2> +0 I'm not informed enough on the issue to really know, but seems fine by me

eprodrom: aaronpk does that make sense to you?

aaronpk: yes

+1

<aaronpk> +1

<bengo> +0

<Loqi> bengo: tantek left you a message 22 minutes ago: can you join us in the telcon 2016-11-01? we need your help resolving some ActivityPub issues interactively so we can get move it along toward PR!

<akuckartz> +0

<wilkie> +0

<sandro> +1 assuming we can do it

<bengo> Time zone fail, I showed up an hour late

eprodrom: any objections?

<cwebber2> bengo, basically we wanted feedback on the things you filed in ActivityPub

RESOLUTION: adopt https://url.spec.whatwg.org/ as URL standard reference for Webmention

<cwebber2> bengo, they are all now Waiting for Commenter :)

<bengo> cwebber2 ya we can just work via github

eprodrom: we are now at 5 minutes past and we are now at the end of the agenda
... i do want to come back to AP, since we were waiting for bengo. would it be woth bringing that up now?

cwebber2: bengo said we can work on github as long as he can help get quick turn-around to get these done by next week

bengo: i will do it right now

<tantek> great!

eprodrom: okay, anything else we need to discuss then?
... thanks for coming everyone

<wilkie> thanks

<eprodrom> Thanks everyone

<tantek> ben_thatmustbeme++ for minuting

<Loqi> ben_thatmustbeme has 177 karma (58 in this channel)

<eprodrom> trackbot, end meeting

Summary of Action Items

Summary of Resolutions

approve https://www.w3.org/wiki/Socialwg/2016-10-18-minutes as minutes for 2016 Oct 18 meeting
publish a new working draft of Social Web Protocols
adopt https://url.spec.whatwg.org/ as URL standard reference for Webmention

Socialwg/2016-10-25-minutes

Social Web Working Group Teleconference

25 Oct 2016

Attendees

Contents

approval of minutes from last meeting

AS2

ActivityPub

PubSub

ActivityPub

SocialWebProtocols

Webmention

Summary of Action Items

Summary of Resolutions