Social Web Working Group Teleconference

13 Oct 2015

Attendees

Present: Arnaud, csarven, rhiaro, aaronpk, shanehudson, sandro, elf-pavlik, kevinmarks, wilkie, eprodrom, jasnell, ben_thatmustbeme, cwebber, tantek, hhalpin, james, tsyesika, wseltzer, akuckartz
Regrets Chair: eprodrom
Scribe: hhalpin

rhiaro: http://rhiaro.github.io/Social-APIs-Brainstorming/socialapi This is an outline of how I understand the different pieces of the SocialAPI and how they fit together or could be considered independantly, which I've started to fill with the equivalent parts of existing specs - as a potential route to convergence.

<cwebber2> and the latest one wasn't this week's meeting up yet ;\

<cwebber2> +q

rhiaro: think its a potential way to move forward

eprodrom: Let's get review this week

it would be good to get a FPWD after our F2F if we have consensus

cwebber2: I did write an email to the call

<eprodrom> https://lists.w3.org/Archives/Public/public-socialweb/2015Oct/0013.html

cwebber2: I was interested in whether or not AS2 should be mandatory or not
... we should address this
... two compoenents:
... 1) One difficulty has been a cycle of conversation between microformats vs Linked Data for the last 15 years

<melvster1> linked data didnt exist 15 years ago

cwebber2: probably not helpful to continue to have those discussions in the WG

<ben_thatmustbeme> melvster1++

I believe it was just RDF

<Loqi> melvster1 has 3 karma

there has been an ongoing argument betwen RDF and microformats obviously.

<kevinmarks> microformats din't either; 12 years maybe

<ben_thatmustbeme> can we get back to the agenda?

scribe: we should specify a deliverable

<tantek> yes KevinMarks 12 years if you count XFN in 2003

scribe: at last call we were worried that there would be no deliverables for client-server API, federation, etc.
... so I think we need to agree whether or not AS2 should be mandatory to implement
... then we can finally move on the API
... but without that level of agreement we are having trouble moving forward

<kevinmarks> +q

eprodrom: What the resolution would look like?

cwebber2: To agree we implement to AS2.0 for whatever is sent to endpoint-to-endpoint
... then we can move forward for other things we are stuck on
... tantek has figured how to map microformats to RDF's type system
... that seems like a good proposal
... then we can figure out whether or not we can move forward

<eprodrom> PROPOSAL: require AS2.0 for Social API and Federation Protocol

<kevinmarks> -1

<aaronpk> uh

<aaronpk> -1

<cwebber2> +1

<wilkie> +1

<wilkie> LOL

cwebber2: It's not other things are not possible, but one is mandatory

+1

<eprodrom> +1

<tsyesika> +1

<melvster1> +0 need to see what AS2.0 is first (mime type will affect existing tooling)

eprodrom: An interesting point about where dynamic is

<csarven> -1

<akuckartz> +0

<tantek> +0 I'd like to improve AS2 more before making it mandatory for anything

<wilkie> practically speaking, a good implementation will have both AS2 and render with microformats. it is practical to have one required though.

kevinmarks: The fundamental disagreement is are we trying to legislate or are we trying to document

<aaronpk> kevinmarks++

<rhiaro_> -1. AS2 could be SHOULD but not MUST

<Loqi> kevinmarks has 173 karma

kevinmarks: what's actually working in the wild
... the point is that the microformat effort has succeeded

<wilkie> well, we could have servers negotiate... do they serve AS2 or just tagged html and then servers will then automatically be required to do both by not requiring one

<wilkie> it's a fun fact of distributed system design

<kevinmarks> microformats have a consistent JSON format for all those languages now

<tantek> we do have JSON explicit in our charter for Social Syntax - just pointing that out in the hopes that we at least don't have to debate that?

<ben_thatmustbeme> JSON is suggested in the charter

<csarven> Charter says JSON, and possible JSON-LD as a convention.

<melvster1> "most people will drop the @context" -- why?

<aaronpk> because it's not actuall necessary to get stuff done

<wilkie> melvster1: because it will probably work without it, and people do that stuff all the time

<ben_thatmustbeme> csarven, i stand corrected

As the author of the charter, JSON is a requirement

JSON-LD was optional

Sorry guys, but that seemed a rather obvious compromise to me

Particularly in terms of developer community size

most modern developers send JSON via HTTP APIs

<ben_thatmustbeme> no, thats a fine compramise to me

See Go, Ruby, Django, etc.

<kevinmarks> documenting is moving things forward

<eprodrom> http://www.w3.org/2013/socialweb/social-wg-charter

<ben_thatmustbeme> i was just being specific about it

<ben_thatmustbeme> i'm fine with JSON

<kevinmarks> mandating things that no-one implements isn't

<csarven> hhalpin Makes perfect sense to me; if JSON, a "convention" needs to be picked, and so JSON-LD is a good candidate.

So, we just need good mappings from mf2 (needs more work but possible) and RDF (JSON-LD solves this, but RDF folks need to stop being so pushy over fine details that most JSON devs don't care about)

<ben_thatmustbeme> hhalpin++

<Loqi> hhalpin has 8 karma

<cwebber2> why have we put all this time into AS2 if it isn't good enough of a structure? :\

<tantek> eprodrom: if we are not comfortable requiring JSON as in the charter, we need to figure out alternative, if we are not comfortable requiring AS2, we need to figure out an alternative

<wilkie> hhalpin++

<Loqi> hhalpin has 9 karma

<tantek> eprodrom: if we can't agree with the charter says, we need to come up with alternatives

<tantek> eprodrom: I'm looking for some guidance

eprodrom: I'm looking for some guidance
... I see a lot of -1s.
... Let's try to resolve this and get a proposal we can accept.

I'm happy to flesh out my proposal over the next week or two in terms of RDF/mf2 interop

<cwebber2> it's fine, I realize it was kind of a bombshell :P

<akuckartz> eprodrom++

<Loqi> eprodrom has 27 karma

That being said, as a reality check, 99% of web devs use JSON, not mf or RDF

<eprodrom> PROPOSAL: accept PTD as an editor's draft for continued development as part of the WG

<ben_thatmustbeme> cwebber2, bombshells are fine but maybe put it on the agenda first

mf has higher uptake in web-pages than any other format

<cwebber2> ben_thatmustbeme: I did, I just messed up and put it on the wrong agenda :P

and RDF has a sizable community as well

<kevinmarks> 100% of web devs use HTML

Type discovery

Yes, but NOT mf2

<cwebber2> so yeah I failed ;p

sorry, but that's true

<tantek> issue-4

<trackbot> issue-4 -- Do we rely on explicit typing or support implicit typing based on explicit property names? -- open

<trackbot> http://www.w3.org/Social/track/issues/4

So we can't force parsing HTML on people who see JSON as easier

<melvster1> kevinmarks++

<Loqi> kevinmarks has 174 karma

<eprodrom> https://indiewebcamp.com/post-type-discovery

<kevinmarks> sure, but claiming that JSON is JSON-LD is more tendentious

<tantek> action-35

<trackbot> action-35 -- Tantek Çelik to Come up with a simple proposal for implicit typing based on property names -- due 2015-02-10 -- PENDINGREVIEW

<trackbot> http://www.w3.org/Social/track/actions/35

however, I agree with the general goal that people who embed data into web-pages using mf seems to work well

<eprodrom> tantek++

<Loqi> tantek has 246 karma

so it should be supported, it has wider uptake

<eprodrom> excellent guessing

tantek: We have implicit typing
... we can get interop in mf2 with AS2

<kevinmarks> no-one is forcing you to parse mf2, we have many parsers for you out there

tantek: its a small, modular draft

my point is *most* people don't use mf2 in modern webdev frameworks. They use HTTP APIs that deliver JSON

<cwebber2> rhiaroif it was the same proposal, but SHOULD instead of MUST for as2

tantek: My request is that the group accept it as editors draft

<cwebber2> rhiaro

however, my second point is the main issue is AS1 and AS2 has little uptake

<cwebber2> would that affect your response btw?

My point is that JSON *itself* has much wider updake than either RDF or mf2

<rhiaro_> cwebber2: I could live with SHOULD

so that should be a common conversion pattern

<cwebber2> rhiaro_: noted

tantek: I would like to keep that mechanism the same

<kevinmarks> no-one is arguing against JSON

furthermore, it's not a small difference: it's several orders of magnitude in terms of uptake

<csarven> :)

<ben_thatmustbeme> i made a quick point that i was corrected on that could have been seen as that kevinmarks

<kevinmarks> mandating the specific JSON-LD vocab that is under draft was what was being proposed

PROPOSAL: Take up tantek's post-type discovery as a Editor's Draft of the WG

<Arnaud> I think this is in keeping with the trend of untyped languages like Javascript with all of its pros and cons :)

<tantek> hhalpin, as an Editor's draft

csarven: The result of ACTION-35 is to change from implicit to explicit approach
... so we have an algorithm to deal with implicit approach

<tantek> https://indiewebcamp.com/post-type-discovery

<Arnaud> the claim "explicit post types that are being abandoned by modern post creation UIs" probably ought to be backed up with some references though

tantek: Accept the document as a work-item
... that's all I'm proposing
... for systems that require explicit typing it expands content they can consumer
... we've had some good experience

<aaronpk> Arnaud, see: facebook and twitter

<kevinmarks> arnaud look at the Medium and Buzzfeed post editors if you want mainstream UI

I am going to note that it appears the mf2 community is the most active implementers

tantek: We've had good feedback

csarven: We would mention that we could say "explicit" or "implicit"

<Arnaud> I'm not questioning the claim, just pointing out that this ought to have a reference, your examples would work

<kevinmarks> instead of a post type, they both have a series of blocks that enable you to post different content types as part fo a post

tantek: I would work with jasnell to see if we could do co-ordinate and browsers tend to do this

csarven: That makes sense

eprodrom: This is the pattern used in micropub
... with POST parameters for intended types of parameters
... correct?

tantek: If you are an AS2 consumer that has explicit types
... the micropub use-case is an explicit one
... one of the implementations is to do this

<melvster1> JSON LD has implicit typing (elf posted some links to the ML)

aaronpk: when I receive a micropub request based on what's in it
... it shows up in different feeds

<cwebber2> melvster1: yeah but the goal here is to help microformats->AS2

<cwebber2> which I think doesn't work with elf's examples

<cwebber2> in that sense it might also be helpful in getting from microformats->rdf

<cwebber2> I think?

<eprodrom> http://indiewebcamp.com/Webmention

eprodrom: There's the types mentioned in a webmention, thats where these properties came from

<rhiaro_> I have my own version to decide internally how to display posts based on the properties: http://rhiaro.co.uk/2015/09/post-type

eprodrom: are we documenting existing behavior?

<ben_thatmustbeme> yes, i do something similar

<ben_thatmustbeme> but this actually helps clear some parts up and makes sense to me

tantek: All sorts of ways of doing responses, but its mainly trying to do things like distinguish between articles and notes

<ben_thatmustbeme> can we just vote on the proposal?

Editors draft does not mean we have consensus, just its a reasonable thing to be working on and within scope

usually in W3C it's a fairly low bar

I would vote for it as it seems useful to a particular community and necessary for interop

<eprodrom> +1

<aaronpk> +1

<ben_thatmustbeme> +1

+1

<cwebber2> +1

<tantek> +1

<rhiaro_> +1

<kevinmarks> +1

<wilkie> +1

<sandro> +1

<akuckartz> +0

<tsyesika> +1

<eprodrom> RESOLVED: Take up tantek's post-type discovery as a Working Draft of the WG

RESOLUTION: Accept Post-type discovery as an editor's draft

eprodrom: Can you move this to the wiki Tantek?

<tantek> will do! thanks everyone

ben: Let's have a flat agenda
... rather than devote too much time to the first topic
... Single items rather than going through actions
... use github issues
... even if it's skipping around topics

<tantek> +1 to flat agenda - allows us more freedom to FIFO different topics and make sure we get to different topics in the order raised / added to agenda

<eprodrom> PROPOSAL: flat agenda where we use bullet points

<rhiaro_> +1

tantek: CSS WG does it this way

<aaronpk> +1

<wilkie> +0

+0

<cwebber2> 0

<Arnaud> I have to admit not to understand what a flat agenda means

<eprodrom> -1

<akuckartz> -0

<eprodrom> q

<Arnaud> it's too bad because I'd be very interested to understand

We could run an experiment and try it once

<Arnaud> but we're out of time

<tantek> Arnaud - see CSS WG telcon agendas as an example

<tantek> posted to www-style

<sandro> -1 I want phone dialin, as much as I like mumble for some things

<ben_thatmustbeme> Arnaud: it would mean this would have had more time than the last minute on the call to discuss this

<Arnaud> to be more explicit: the problem seems to be about time allocation

<sandro> but we could try one of the other phone dialin options

any objections to mumble?

<Arnaud> I don't know that the format of the agenda is the problem

<cwebber2> so, sounds like objections. ok!

What we need is mumble + phone dial-in :)

<akuckartz> +1 to using open source compatible solution

<cwebber2> but

<cwebber2> I think we still need to handle the international phone version :P

Although notes non-US folks can't dial-in as easily

<tantek> -1 I want phone dialin also like sandro

<eprodrom> agenda item for next week?

eprodrom: lets revisit next week

<ben_thatmustbeme> can it be the top of the agenda next week?

<ben_thatmustbeme> :P

<tantek> hhalpin: I'm using Google Hangouts on iOS over wifi to make a "phone call"

<kevinmarks> the workaround for no-us is to use google hangouts which does require an account, but has free us calling

eprodrom: meeting adjourned, talk to you next week

<tantek> hhalpin++ for minuting!

trackbot, end meeting

Socialwg/2015-10-13-minutes

Social Web Working Group Teleconference

13 Oct 2015

Attendees

Contents

Approval of minutes from last week

Discussion of F2F

AS 2.0

Social API

Type discovery