Social Web Working Group Teleconference
28 Apr 2015
See also: IRC log
- Ann, +1.408.335.aaaa, Sandro, Arnaud, jasnell, elf-pavlik, +1.617.247.aabb, aaronpk, tantek, ben_thatmustbeme, cwebber2, eprodrom, rhiaro, aaaa, Tsyesika, bblfish, +1.314.777.aacc, harry, Ann_Bassetti
- AnnB, harry
<AnnB> scribe: AnnB
<eprodrom> scribenick: AnnB
<eprodrom> scribe: AnnB
approve minutes from 21 April
<tantek> +1 LGTM
<eprodrom> PROPOSED: approve minutes ffor 21 April
<eprodrom> RESOLVED: approve minutes ffor 21 April
resolved to approve
Evan: NO regular telecon next week, as F2F will be in Paris
Evan: next telecon will be 12 May
... agenda for F2F, next week (4 and 5 May) ...
... do we have outstanding issues for that meeting?
Arnaud: not that I know
... but, what about food?
Sandro: no food provided, but Paris does not lack for cafes
Harry: location is INRIA's office
<Loqi> Jsnell made 2 edits to Socialwg/2015-04-28 https://www.w3.org/wiki/index.php?diff=83885&oldid=83873
<Loqi> Pelf made 1 edit to Socialwg/2015-04-28 https://www.w3.org/wiki/index.php?diff=83886&oldid=83885
Harry: south side of Paris
<elf-pavlik> +1 securing food :)
Harry: will put up signs to room
... probably coffee
... no food
Sandro: looking at email from <someone> ..
... breakfast can be delivered for a price .. not lunch
... one must pay on their own
Harry: yes, will be there
<tantek> +1 glad to hear you'll be there Harry
Evan: great participation; ~19 people
... worse things than having to go out to lunch in Paris
<harry> W3C should support its Working Groups. I'll ask MIT re coffee and food today, no help from ERCIM despite EC funding this effort but I can ask again.
<KevinMarks> go out to lunch and breakfast, people
Sandro: if anyoen can fund food, send me email pronto
Evan: agenda filling out
Evan: chairs need to do final organizing
<harry> However, yes worse fates than eating lunch or having coffee in Paris.
Evan: last F2F we ran out of time on some important items
... so let's prioritize this time
... Outstanding Issues and Actions
Evan: any of those ready to be closed?
<harry> Let's not walk through them one by one
Evan: <no one speaks>
<harry> if there are no updates - but everyone should check.
Evan: Activity Streams vs LDP Collection and paging structure raised this week
<trackbot> issue-37 -- LDP & AS2.0 paging alignment -- raised
<tantek> -0 (don't understand why it's relevant)
<harry> The larger question is probably if we have to chose between backwards compatibility with ActivityStreams and LDP, which do we do?
Evan: I'd like to propose this Issue be Opened
jasnell: -0 because we don't have concrete proposal on table
<harry> My guess it would depend it on the number of ActivityStreams 1.0 deployments out there vs. number of LDP deployments.
jasnell: LDP paging relatively new; no implementations
<KevinMarks> there's a McDonalds across the street http://imgur.com/KvOwzg1 ;)
<harry> So answering that empirically is probably right way IMHO
<tantek> +1 to jasnell's reasoning
jasnell: opening it is maybe OK, but ..
<sandro> +1 jasnell let's wait until there's a concrete proposal for changing AS2
<bblfish> I added SoLiD to the https://www.w3.org/wiki/Socialwg/2015-05-04#Evaluation_of_API_Candidates
Evan: IMO if we have issue that's pertinent, better to open it in order to have discussion
... even if we need to close it
<KevinMarks> this would eb a better choice - thanks foursquare: https://foursquare.com/v/la-butte-aux-piafs/51d2d6d8498e1573c8e26a94
bblfish: question of seeking compromise for both LD and JSON communities
... I need to understand more about JSON-LD, for example
<Arnaud> I do like the way LDP Paging separates paging from the data by using headers but this is a fairly significant change
<Loqi> Eprodrom made 1 edit to Socialwg/2015-05-04 https://www.w3.org/wiki/index.php?diff=83887&oldid=83883
<Loqi> Aboyet made 2 edits to Socialwg/2015-05-04 https://www.w3.org/wiki/index.php?diff=83889&oldid=83887
<Loqi> Bblfish made 1 edit to Socialwg/2015-05-04 https://www.w3.org/wiki/index.php?diff=83890&oldid=83889
Evan: question: is this an issue for us, if a SoLiD API is one of the outputs of this grouip
<harry> My guess would be some parts of SoLID might go into the Social API or ActivityStreams 2.0, but not all of it.
<harry> (There is BTW, no such thing as a false graph in RDF)
bblfish: point is not to tie AS to LDP at this point, but to figure out the minimal RDF
<harry> And "really doing RDF" is not a requirement in this WG BTW
bblfish: not inferencing, etc
... just merging 2 graphs ... which I think is done correctly in LDP paging spec
<harry> There are probably a number of different ways to define "really doing RDF" :)
<tantek> harry - exactly - I don't understand this at all
<jasnell> btw, look at http://dandus.mybluemix.net/r/ for an example of AS2 + paging links in the HTTP headers
bblfish: can think about how could that be done in JSON-LD
<harry> I don't think it's really important but happy for it to be an open issue.
<sandro> I thinks it's time for the chairs to re-assert their authority to decide which issues to open, because they're worth the time of the WG
<Arnaud> every issue we open adds to our workload
<tantek> sandro agreed
Evan: asks queue to focus on question about opening this issue, NOT the technical merits of this and that
<Zakim> elf-pavlik, you wanted to discuss do we need concrete proposal to OPEN raised issue?
<tantek> I'm still not convinced, I've only heard a bunch of plumbing-specific reasoning.
<tantek> hence still -0
<jasnell> ah... nevermind, looks like that deployed version has the headers switched off ;-)
<sandro> -0.5 to opening this issue. Let's not make up work for ourselves. If there's a problem, THEN raise the issue.
<harry> Anyways, re the giant collection of possibly unused technolgies we could discuss, l suggest that someone just take homework looking for implementations.
<tantek> "as a placeholder"?!? no that's a horrible idea
<tantek> sandro - that's a good way of puttin git
<harry> So, for Webfinger, does anyone still support it?
elf-pavlik: I support opening the issue
<tantek> changing my answer to -0.5
<harry> Ditto LDP paging
<eprodrom> harry, pump.io does
<aaronpk> agreed with "if there's a problem, then raise the issue"
<harry> Yep, exactly.
elf-pavlik: do not need to have clear resolution before opening
<harry> That's the kinds of data-points we need to make decisions.
jasnell: back to RDF and merging of graphs..
<tantek> this seems like a "might be a problem" issue, not an *actual* issue - thus I'm starting to lean towards rejecting
<harry> I am not sure if we want to get deluged in open issues.
jasnell: 1) we're not dealing with RDF
... yes, someone might use RDF for modeling, but not a requirement
<harry> So at some point, we might want to tighten raising issues. Typically, I'm happy to give the Editor discretion in opening most issues.
jasnell: but what is the issue? not clear
<tantek> agreed with jasnell
jasnell: need better definition
... what does not work in current model?
... in JSON what does not work
bblfish: hard to answer right now; could do on email
<tantek> many of us do not care about doing RDF
bblfish: those who are doing RDF could use it
<tantek> so it is inaccurate to express "we are doing RDF" for the group
<harry> RSS 1.1 was not exactly an amazing success story :)
<KevinMarks> we're doing pure JSON with a way for RDF to be created from it
bblfish: trying to find compromise that satisfies multiple communities
<KevinMarks> RSS 3 got more adoption than RSS 1.1
<tantek> has this affected any implementations? or spec language?
bblfish: <scribe missed who / what .. divergent ways of modeling>
<tantek> AnnB - that was bblfish
<Arnaud> I think this is primarily a design style issue
<harry> I agree we should respect the divergence if it has implementation impact (I don't count ontology differences, given one side doesn't use one, as an implementation impact)
<eprodrom> PROPOSED: open issue 37
<trackbot> issue-37 -- LDP & AS2.0 paging alignment -- raised
<tantek> -0.5 per Sandro's reasoning, and jasnell 's reasoning
yes, tantek .. I think I started with bblfish at beginning of that sequence
<sandro> -0 it's probably a waste of time at this point
<aaronpk> -0 per sandro's reasoning
<tantek> group's time is limited, and issues not affecting a specific implementation or spec should not be opened
<tantek> let's get better at raising *specific* concrete issues
<harry> scribenick: harry
<bblfish> ok I'll do
<bblfish> ok harry
<Loqi> harry has 11 karma
<ben_thatmustbeme> scribe: harry
evanp: ISSUE opened.
<jasnell> I'm still -0
<tantek> either way, no difference, no objection.
<AnnB> CAN SOMEONE TYPE>>> KEYBOARD WENT BAD (SORRY)
<eprodrom> AnnB: harry is going to step in
<eprodrom> scribe: harry
eprodrom: Deadline for today to get Social API candidates ready for review
<bblfish> RESOLVED: Issue37 opened
<wilkie> ooh ok
eprodrom: the intention was to have SoLID, Pump.io, and Micropub ready to discuss
... any updates?
<Tsyesika> we do, it's https://w3c-social.github.io/activitypump/
<cwebber2> eprodrom: Tsyesika doesn't have a working mic today fyi
<Tsyesika> (sorry i don't have my mic)
<Loqi> Tsyesika has 7 karma
The link to the Pump.io spec is old in the wiki, update real quick?
evprodrom: We have a working review
<Loqi> Tsyesika has 8 karma
eprodrom: Let's set-up for evaluation of API candidates - do we have a micropub-based proposal?
<Loqi> aaronpk has 786 karma
aaronpk: I'll do another pass on it today but its ready for review
sandro: SoLID is not as polished as we like, but it's feature complete
eprodrom: All up and ready to discuss
... Linked to Hydra and Linked Data Fragments
<elf-pavlik> i'll demo at least Hydra
eprodrom: note that neither of those are there.
... so part of our resolution was to stick to these three
<rhiaro> I thought sticking with the list was until another demo came forward?
eprodrom: unless it's a real 'going to blow us away' not sure if we should leave these open
<Tsyesika> I thought new candidates could be accepted later if they could demo?
So, let's remove them from candidates
<Arnaud> I agree with Evan's take on this
elf-pavlik: We don't have a query langugae
<Arnaud> we need to further narrow the scope, not widen it :)
elf-pavlik: they are not competing not complimentary
eprodrom: Remove them from list, put them on another section
... if time for them being discussed, let's do that.
<Zakim> tantek, you wanted to note that we did leave the door open at the last f2f for more demos as a way to introduce additional API candidates.
+1 narrow focus
<cwebber2> ah, tantek is right, I forgot about that
tantek: We did leave it open, and we would need a demo of specific user-stories
My opinion is that we'll add them to candidates at f2f if we cover user-stories with those technologies.
scribe: just make sure they actually cover user-stories
<elf-pavlik> Additional Candidates
<elf-pavlik> (if time allows and have demos for existing user stories)
eprodrom: The most important thing is for people remotely participating and to read over the documents
... everyone should have some notes on each document
... be prepared to address them.
... do we have for each one for notes to be submitted?
<Tsyesika> open issues? :)
The others are on github
<tantek> or you could post on your own site feedback about Micropub and send a webmention ;)
notes we'll be moving ActivityStreams and its testing to W3C github's account.
<elf-pavlik> could everyone add those link to F2F wiki?
<aaronpk> i will definitely read those :)
<Loqi> Bblfish made 1 edit to Socialwg/2015-05-04 https://www.w3.org/wiki/index.php?diff=83891&oldid=83890
<Loqi> Pelf made 1 edit to Socialwg/2015-05-04 https://www.w3.org/wiki/index.php?diff=83892&oldid=83891
eprodrom: any other API discussions?
... thanks to everyone who has taken up agenda
... lots of agenda
<AnnB> not really >>
<AnnB> getting weird chars
<AnnB> i can type some< but wonky
jasnell: We have W3C github accounts set up by harry
<AnnB> e>g> can"t type slash
jasnell: IBM gave us approvals, we should have test-suite on github
... J.P is moving it out
... we will look to community to fill it out
Does J.P. have a github ID?
Or does it go via you?
(either is fine)
<tantek> come back AnnB!
Simplifying audience we're addressing
<Tsyesika> i can't hear ><.<
<Tsyesika> * >.<
jasnell: Can we simplify the issues?
Just noting that *anyone* can do pull requests
but that W3C Social editors are in charge of approving push/merge
jasnell: mismatch between icons and the rel attribute
Zkaim, mute me
scribe: for the most part, I'm +0
scribe: I see it as a simplifications
... parallels with atom link and links in html
... point it can be confusing
... difference between property name
... remove 'rel'
... use the property name as link relationship
eprodrom: We will have other encodings using IETF link relationships
... HTTP headers, etc.
<tantek> HTML5 does not use IETF link relations, FWIW
<elf-pavlik> HTML5 uses microformats link relations
<tantek> W3C formats have settled on microformats link relations
<tantek> and informally last week I asked Mark Nottingham if HTTP2 could use microformats link relations and he said sure
<jasnell> the status quo is that "rel" must be compliant to both IETF link relations and HTML5 link relations
eprodrom: the whole point is to have 'rel'
<KevinMarks> cite for evan: http://www.w3.org/TR/html5/links.html#other-link-types
eprodrom: I see no reason to remove
<jasnell> I'm -+0 to the proposal to remove "rel"
<tantek> I'm still not sure how as:Link works, but am a moderate fan of rel
<tantek> note that link relations predate Atom - started in HTML long before that
bblfish: If you look at link relationships, on what element does the title come?
<jasnell> I'm definitely -1 on removing as:Link
bblfish: so as I argued that if you do a mapping from link relationships to an ontology where I modelled it (Atom-OWL)
... there's not many ways you can do this correctly
... map to RDF Reification
... the link relationship is just a reified relation
... so you can transform it into a relationship
... so its a big simplificiation
<tantek> wow I have no idea how to process the thread at https://github.com/jasnell/w3c-socialwg-activitystreams/pull/100 nor what problem it is trying to solve.
<tantek> is that what happens when you use JSON instead of HTML to do an <a href rel> ?
should probably be a single way to do rel, but I think we have that in terms of as:Link
<tantek> is anyone consuming as:Link ?
<jasnell> -1 to removing as:Link
eprodrom: Is there a way to remove as:Link with less pressure?
<aaronpk> what is as:Link for?
<elf-pavlik> +1 removing as:Link
-1 to removing as:Link
<tantek> if pump.io is using as:Link, that's a strong argument to keeping it
jasnell: as:Link sets up a qualified linked relationship
... I can include the note object
<cwebber2> 0 on removing as:link ;)
jasnell: right there or I can describe a link to that note
... the difference is that the title may be different
<KevinMarks> I have just been adding mf2 rel parsing to universal feed parser, so that link's in the body with rels end up in the parsed collection of lnks
jasnell: two different references have different metadatas
jasnell: maybe what is described is reference to object
<Loqi> tantek has 182 karma
jasnell: if you break the document into triples
... if you don't have as:Link, metadata steps all over each other
... Description of reference
<KevinMarks> I have 3 dogs with me but they aren't barking and I'm muted
<tantek> aaronpk: does that mean the URL provides the canonical information?
jasnell: so you don't want to duplicate that URL data on all the types
aaronpk: Would you have to duplicate?
jasnell: it's the same as links in Atom
... same as anchor tag
jasnell: as:Link allows you describe the reference
aaronpk: it allows you to describe the reference
jasnell: I'm against removing as:Link
<KevinMarks> if you're having links, rels clarify them
tantek: Keeping rel and Link would allow you to keep easier compatibility with HTML
... more compatibility with what is on web
bblfish: This is a modelling discussion
<eprodrom> This will be the end of our discussion, btw
<jasnell> I'm fine with keeping rel. But some folks have problems with it so I put the proposal on the table to see where the consensus is
bblfish: since it's just a reificiation, then you lose a lot from leaving it out
bblfish: I propose one can do it
... see graphics on that page
<elf-pavlik> I can prepare clear diagrams and exmaples of the issue for F2F (or compile existing ones spread across mailing list and gh issue)
bblfish: relationships to alternates of that
<tantek> what's the problem that the example is trying to solve?
bblfish: blank node, too complex for a telecon
tantek, I think bblfish is trying to figure out how to map it in a way he likes to an ontology
<tantek> what's the real world problem that needs discussing blank nodes?
he is assuming you are using RDF. The 'real world problem' is unclear to me as well
<tantek> harry - I'm having difficulty understanding all the architecture diagrams unrelated to a specific real world publishing problem.
<elf-pavlik> including issues with as:Link instances included as items of as:Collection
<tantek> but I'll keep reading
In general, reificiation is not even recommended in RDF.
<scribe> scribenick: AnnB
<harry> scribe: AnnB
bblfish: I think I'm OK w removing the REL, but should be worked offline
Evan: nearing end of meeting; how to proceed?
<tantek> I'd rather keep compat with the simple HTML model of publishing and hyperlinks.
Evan: suggest leaving it up to editor to continue
... proposals on guidance to editor?
<tantek> PROPOSAL: keep AS2 compat with HTML <a href rel type>
Evan: other proposals?
<tantek> (based on compat with existing publishing practices)
<bblfish> I am not sure I understand that proposal
Evan: not closing this; just giving guidance
Evan: it's up to the editor
<bblfish> I am for working on these arguments more closely.
Evan: asks elf-pavlik and bblfish for opinions
<elf-pavlik> we need to develop little more common understanding
<harry> i.e. rejecting elf-pavlik's suggestion to drop
Evan: on proposal as written by tantek
<elf-pavlik> i have impression that some people might have not read or understand http://www.w3.org/TR/activitystreams-core/#link
Evan: <someone> points out that is status quo
bblfish: no opinion; just think the discussion isn't over
Evan: I agree re: discussion; just trying to wrap up for today
... looks like all +1 and 0 ..
<elf-pavlik> -0 without as:Link and as:rel we still stay compat with HTML
<eprodrom> RESOLVED: keep AS2 compat with HTML <a href rel type>
<harry> Seems keeping compatibility with Atom makes sense to me
Evan: thanks everyone
<harry> chairs - time to go over f2f logistics
<eprodrom> Zakim who is on the call?
<cwebber2> thanks all
<harry> trackbot, end meeting
<Arnaud> that won't kick us out