#social 2015-07-06

2015-07-06 UTC
jasnell joined the channel
#
melvster this page is nice work : http://www.w3.org/wiki/Socialwg/Social_API/User_stories/Groupings
KevinMarks, jasnell, bblfish, tilgovi and the_frey joined the channel
#
melvster are any of the user stories implemented yet?
#
elf-pavlik melvster, https://github.com/w3c-social/social-ucr#proposed-implementations
#
melvster elf-pavlik: does that mean "no"?
#
elf-pavlik well, Indie Web has live deploymens of basic ones like 'posting a note', 'posting a photo' etc.
#
melvster elf-pavlik: it's really not clear what is implemented and what is brainstorming
#
melvster seems like <5% of the use cases are implemented?
#
melvster but im guessing
jaywink joined the channel
#
elf-pavlik melvster, if you would like to step up to help with tracking implementations i believe everyone would appreciate it
#
melvster elf-pavlik: sorry, if i did that I'd have to drop something else (hint: like actually implementing stuff) ... it's hard to know what's been dont and not dont, is all im saying
#
melvster s/dont/done
#
elf-pavlik well, enough if you keep track on what you implemented in a way the whole group can easily find it and aggregate with what other people have already implemented
#
elf-pavlik i hope i can have demo of WebID-TLS authentication with OAuth Bearer Tokens authorization https://github.com/w3c-social/Social-APIs-Brainstorming#bearer-tokens
#
melvster well that would be useful
#
elf-pavlik i think this can align SoLiD and other two even closer
#
melvster yes
#
melvster you are correct
#
elf-pavlik it also nicely separates resource server from authorization server and we look at indieauth that person can also specify service used for verifying identity
#
elf-pavlik which in turn can support WebID-TLS, Persona, OpenID Connect etc.
#
melvster agreed, very valuable
#
elf-pavlik i still need to chew more on OAuth confidential vs. public clients
#
elf-pavlik since clients with backend (e.g. most micropub clients as of now) clasify ans confidential ones, while those running fully in a browser (e.g. most SoLiD clients as of now) clasify as pluibic ones ...
#
elf-pavlik http://tools.ietf.org/html/rfc6749#section-2.1
#
elf-pavlik s/pluibic/public/
#
elf-pavlik needs to run soon
#
melvster good idea
bblfish and Arnaud joined the channel
#
Loqi Alehors made 1 edit to [[Socialwg/2015-06-30-minutes]] https://www.w3.org/wiki/index.php?diff=84923&oldid=84900
#
Loqi Alehors made 2 edits to [[Socialwg]] https://www.w3.org/wiki/index.php?diff=84925&oldid=84810
#
Loqi Alehors made 1 edit to [[Socialwg/2015-07-07]] https://www.w3.org/wiki/index.php?diff=84926&oldid=0
bblfish joined the channel
#
Loqi Alehors made 3 edits to [[Socialwg/2015-07-07]] https://www.w3.org/wiki/index.php?diff=84930&oldid=84926
#
Loqi Alehors made 1 edit to [[Socialwg/2015-07-07]] https://www.w3.org/wiki/index.php?diff=84931&oldid=84930
jasnell joined the channel
#
melvster agenda: https://www.w3.org/wiki/Socialwg/2015-07-07
tilgovi joined the channel
#
cwebber2 hello #social
#
ben_thatmustbeme hello cwebber2
#
cwebber2 hi ben_thatmustbeme
#
ben_thatmustbeme lots of AS2 discussion for tomorrow thus far
#
elf-pavlik ben_thatmustbeme, would you like to try https://elf-pavlik.github.io/unmpc/?url=https://phubble.tuxed.net/w3c-social/
#
elf-pavlik and https://elf-pavlik.github.io/unmpc/?url=https://phubble.tuxed.net/alice-b-day/
#
elf-pavlik they both use this data for very primitive way of defining ACL
#
elf-pavlik oops! i noticed nasty bug :( to see alice-b-day secret wall one can't easily sign in (get auhorization token) first ...
#
ben_thatmustbeme elf-pavlik: i don't even know what i'm supposed to do with that
tilgovi joined the channel
#
elf-pavlik you can post to shared spaces, a public one https://phubble.tuxed.net/w3c-social/ and a secret one https://phubble.tuxed.net/alice-b-day/
#
elf-pavlik you can try directly or via unmpc, both works similar now
#
elf-pavlik to my surprise we possibly didn't capture common use cases like https://twitter.com/SocialWebWG
#
melvster can anyone explain to me how inboxes and outboxes work?
melvster joined the channel
#
ben_thatmustbeme ahh, elf-pavlik there it goes, the link wasn't working before, all i got was the 2 input boxes and the sign in one was grayed out
#
ben_thatmustbeme melvster: regarding user stories, have you seen
#
ben_thatmustbeme https://www.w3.org/wiki/Socialwg/Social_API/Sorting_user_stories ?
#
melvster ben_thatmustbeme: thanks, i might have seen this before, but its really hard to find going from the first wiki, in fact the user stories themselves are hard to find ... it's pretty confusing, but that's a great page, now that I see it ...
#
melvster ben_thatmustbeme: but it still doesnt tell me if any of them are implemented, or going to be implemented
#
ben_thatmustbeme i need to review if we resolved on it or not, but i THINK that we said the 7 entirely positive ones are approved
#
ben_thatmustbeme in addition to SWAT0
#
melvster ben_thatmustbeme: OK, thanks. I've given up on swat0. I dont think it's possible without missing information. I'm waiting to see how others do it, to understand better.
#
aaronpk melvster: re: your last email, the concept of "friends" is not actually critical to this, since even Twitter doesn't have that concept
#
ben_thatmustbeme true, its rather difficult to know that. Honestly many of the user stories have parts implemented in indieweb
#
aaronpk one-directional following is enough
#
ben_thatmustbeme thats why i don't go by strictly counts of +1's etc
#
melvster imho, microblogging is just one application of the social web, the social web being a bigger thing, I do think friends are fundamental to social ... id be willing to bet most people outside of this group would probably agree ...
#
ben_thatmustbeme a lot of people put things like "+1 but" or "+1 on the condition that"
#
aaronpk a lot of things are fundamental, but not required for an initial API
#
ben_thatmustbeme can't require friends either
#
ben_thatmustbeme all of twitter would say so
#
ben_thatmustbeme and G+
#
melvster yes and all of facebook would say you do need friends :)
#
aaronpk also I doubt any corporate networks want to have a "friend" concept
#
ben_thatmustbeme and actually FB has moved to that model of "friending but ignoring all their content"
#
aaronpk way too loaded of a term to use in that setting
#
aaronpk yeah I much prefer the "friending" on facebook not being tied to following people's content
#
ben_thatmustbeme which is interesting actually. friending and following are two seperate things.
#
melvster yes
#
melvster makes implementing things really confusing
#
ben_thatmustbeme but again, if we are looking for a small set of common cases, friending isn't that popular
#
aaronpk exactly
#
melvster friending got 10 +1's and 3 +0's
#
melvster no - 's
#
melvster even inbox got one -1
#
melvster and that's considered 'all positive'
#
ben_thatmustbeme melvster: read those again, MANY of those +1s are saying +1 but it should be one-way
#
ben_thatmustbeme no, there are 0's that makes it not "all positive"
#
melvster +0 is different from 0
#
ben_thatmustbeme yes, and there is a 0 vote
#
melvster only 1
#
ben_thatmustbeme 0 Unsure both how valuable it is and if it's really something to do at the API level. Also does Tammy unfriending Jeremy stop Jeremy following Tammy (so he'd get objects shared to Public or those Following her?) Jessica Tallon (talk) 18:43, 23 February 2015 (UTC)
#
melvster and inbox has a -1 but is still considered 'all positive'
#
melvster following is already a separate use case
#
ben_thatmustbeme the non-clarity of the use-case is then a great reason to down vote it
#
ben_thatmustbeme remember these are votes on the USER STORY, not on the functionality
#
ben_thatmustbeme these are not votes to include / not include functionality. many were downvoted because the user story was not clear / specific
#
ben_thatmustbeme does not mean there is not any room to improve them and get rid of downvotes
#
melvster it's inconsistent
#
melvster swat0 isnt easily implementable or understandable, it's been around 5 years and no one has implemented it, yet it's somehow the gold standard
#
melvster caveat aaronpk: has come by far the closest
#
aaronpk i thought there were actual implementations a long time ago? status.net?
#
ben_thatmustbeme an acid test is supposed to be difficult to implement
#
ben_thatmustbeme is it not
#
melvster no idea
#
ben_thatmustbeme how long did it take browsers to pass ACID3?
#
ben_thatmustbeme hell chrome still gets only a 98% here
the_frey joined the channel
#
ben_thatmustbeme actually... i think we can do SWAT0 now. I need to check my reference of tagging. but i can post a photo in mobilepub and tag it with aaronpk i think. i might need to add that field
#
aaronpk we have a pretty thorough writeup here https://indiewebcamp.com/SWAT0
#
melvster direct messaging got only 9 +1s
#
melvster friending got 10 +1s
#
ben_thatmustbeme melvster: # of +1s was originally said not important
#
ben_thatmustbeme there were so many we said you don't need to vote on all
#
melvster well i would have voted for the friending one
#
melvster but it would have made no difference
#
melvster because of the +0
#
melvster did people know that +0 would essentially mean a use case was not approved?
#
melvster and some of these use cases have dependencies on other use cases ... and those ones are not approved ... we need a dependency tree ...
#
ben_thatmustbeme okay, so nothing has been "disapproved yet" they just have not been approved YET
#
ben_thatmustbeme anything that has +0s we want to look at closer and see "why not a +1" before we approve
#
aaronpk the goal really is not to approve a whole bunch of user stories
#
ben_thatmustbeme there is no dependance on other user stories, only on functionality they describe
#
melvster ben_thatmustbeme: some require the inbox
#
ben_thatmustbeme thats fine, they require an "inbox" functionality, something i'm not sure it really clearly specified. but the "Inbox" user story in one of an infinite set of user stories that can be used to describe such a feature
#
ben_thatmustbeme also, maybe "inbox" is not needed as a user story at all as its existance is specified by there being a user story that requires it
#
melvster so how do we implement inbox then?
#
melvster does a user have one or can they have many?
#
melvster this is a really important implementation detail
#
ben_thatmustbeme so then explore, try things out
#
ben_thatmustbeme see what makes sense and come back to us on that
#
melvster try what out? i need to know the relationship one one or one many BEFORE implementing it
#
melvster these things need to be specified prior to an implementation
#
aaronpk hehe
#
ben_thatmustbeme this isn't "here's your assignment" type of stuff. its exploratory coding.
#
aaronpk not really
#
aaronpk see: indiewebcamp
#
melvster yes but an inbox isnt defined anywhere
#
melvster so one one knows what they are supposed to be implementing
#
melvster someone might come along and say, 'i used gmail as my inbox'
#
melvster then what?
#
ben_thatmustbeme melvster: nothing is fully defined yet. we are in the process of defining them. the user stories are something to give us a direction of user interaction, they are not specs, should not be treated as specs
#
ben_thatmustbeme if you want to work with indiewebcamp we don't really use inboxes, just send webmentions to pages. Pump.io does use them, they specify them. or make your own and see how that works
#
ben_thatmustbeme test out various implementations
#
melvster ben_thatmustbeme: so is indieweb going to implement inboxes then?
#
ben_thatmustbeme based on our implementations its not needed
#
melvster ben_thatmustbeme: is indieweb actually planning to implement the social web api?
#
aaronpk or rather, the webmention endpoint acts sufficiently like an inbox
#
ben_thatmustbeme true, webmention endpoints are basically indieweb equivalent of them
#
aaronpk melvster: there is not a single "indieweb" as you speak of. there are many people implementing many different things that work together.
#
melvster wait, i thought webmention were ENDPOINTS not inboxes?
#
aaronpk whether someone implements the w3c social API will be based on whether there is sufficient motivation for each person to go to that effort
#
ben_thatmustbeme whats the difference. it isn't specified
#
melvster ben_thatmustbeme : but then you could say anything is an inbox?
#
melvster which means it most be one many
#
melvster or actually many to many
#
melvster because one inbox could be for a whole family too, right?
#
ben_thatmustbeme until we actually create a spec, sure.
#
ben_thatmustbeme i think a single "inbox" for a whole social network could work
#
ben_thatmustbeme would allow bridging in to silos
#
ben_thatmustbeme if a webmention endpoint is an indieweb version of an inbox, then thats exactly what webmention.io is
#
melvster ben_thatmustbeme: but once you advertize an inbox (ie like email) what's to stop tons of spam filling it?
#
melvster in the old days people used to hide their email addresses
#
ben_thatmustbeme good question to ask. spam is definitely a problem that needs to be addressed, but not sure its within the scope of the group. tantek had proposed a very interesting way to severely cripple spam in webmention with "vouch"
#
ben_thatmustbeme which is actually running on my site already, aaron's too i believe
#
ben_thatmustbeme also spam filtering is pretty good on its own
#
melvster ben_thatmustbeme: so you vouch for all your friends?
#
aaronpk nothing about Vouch involves friends
#
aaronpk it's worth reading http://indiewebcamp.com/vouch
#
melvster so friends of friends then
#
aaronpk no
#
aaronpk nothing about vouch implies any sort of "friend" relationship
#
melvster what's 449?
#
aaronpk http://indiewebcamp.com/vouch#Is_vouch_saying_you_know_me_from
#
aaronpk "HTTP 449 Retry With"
#
aaronpk https://msdn.microsoft.com/en-us/library/dd891478.aspx
#
melvster 'A Microsoft extension.'
#
aaronpk seemed appropriate
#
melvster is it right to say its non standard then?
#
melvster aaronpk: why not 401?
#
aaronpk because we don't use the HTTP Authorization header
#
aaronpk "The request requires user authentication. The response MUST include a WWW-Authenticate header field" http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
#
melvster but surely you're trying to show the request is 'authentic'
#
aaronpk no
#
aaronpk just that it is vouched for by someone else
#
melvster id by providing a token a 3rd party gave you
#
aaronpk no tokens involved
#
melvster aaronpk: i think crypto currencies could be a good solution to this problem
#
aaronpk that has been discussed as well
#
melvster then you could use 402
#
aaronpk https://indiewebcamp.com/Webmention#proof-of-work
#
melvster aaronpk: consider a postage stamp ... you send it and pay for that ... but now imagine the end user sees your postage stamp, sees it's NOT spam, and doesnt cash it ... but you're thru the spam filter ... now a spammer would spend money on the stamp, but if it's spam the end user cashes it, providing a disincentive to spam ...
#
aaronpk people still send junk physical mail all the time
bblfish joined the channel
#
melvster but if it was expensive they would stop
#
melvster aaropk: proof of work like that is a bad idea, because you burn cpu cycles for no reason
#
aaronpk you are welcome to propose another solution
#
aaronpk or even better, implement it so that someone else can implement the other half
#
melvster yes i could do that
#
melvster using http 402
#
melvster "payment required"
#
melvster it doesnt say the end user will actually CASH OUT that payment
#
aaronpk well, "payment required" to me implies that I will not have that money after sending it
#
aaronpk but that isn't really defined by HTTP, so that's just my interpretation
#
melvster aaronpk: im currently speccing out http 402 ... just as vouching for someone doesnt imply a friend, payment required doesnt mean it will cost you something, payment is required for a beer, but you might get one free ...
#
melvster (as in beer)
#
rhiaro reads backlog
#
aaronpk that is a sublte distinction that I'm sure could be argued but I have no interest in doing so, nor do I think it is particularly relevant to the implementation
#
aaronpk regardless, I would be interested to see what that looks like speced out and implemented
#
rhiaro I'd just like to throw in that I agree a webmention endpoint is equivalent to an inbox endpoint (as specified by ActivityPump). The difference being that AP also specifies some things a server should do once it receives something at the inbox endpoint, whereas webmention doesn't. Note, AP doesn't have a good solution to spam either
#
melvster aaronpk: it's not really subtle ... consider a night club, there may be a cover fee, but if you're a VIP because you know the owner you get in free ... this is EXACTLY like a vouch ...
#
melvster a 402 keeps people out in general without paying, but some will not need to
#
melvster it would be stupid to charge the owner of the nightclub a cover fee, for example
#
aaronpk like I said, I'd love to see this written out
#
aaronpk feel free to brainstorm on the indiewebcamp wiki
#
melvster aaronpk: ok will do, ive actually implemented 402 locally already
#
rhiaro melvster, "proof of work like that is a bad idea, because you burn cpu cycles for no reason" - the reason is to find implementations that work well in practice (rather than just in theory), no?
#
melvster rhiaro: i didnt mean to imply there's not a reason for doing it, i meant that you can do the same thing by reusable proof of work, rather than destroyable proof of work, which would use less electricity
#
aaronpk the idea of "you can send me something, and I will take your money only if I don't want it" is interesting
#
aaronpk the sender has to decide how much they are willing to risk losing on the communication
#
ben_thatmustbeme would "accidently" just accept all money anyway :P
#
melvster yes!
#
rhiaro oh sorry melvster, I thought you were talking to proof of work as a general concept, not just for webmention vouch
#
aaronpk ben_thatmustbeme: that would mean I would quickly stop sending you things :)
#
ben_thatmustbeme haha
#
melvster this ^^
#
ben_thatmustbeme also a bit of limitation of automating things then
#
melvster it's like buying someone a beer and they dont get you one back
#
ben_thatmustbeme #tooManySimilies
#
ben_thatmustbeme actually the party similie was exactly how vouch was named
#
melvster west coast hipsters :)
#
melvster ok ill implement this and write it up
#
melvster i actually use it every day
#
melvster only problem will be then that vouching for someone could be bought ...
#
aaronpk Vouch is meant to solve the spam problem, not prevent targeted attacks
#
melvster aaronpk: so do you have different inboxes say, for public, for private, for friends etc. ?
#
aaronpk i only have one webmention endpoint
#
melvster aaronpk: but is it possible, will it be needed?
#
aaronpk I can easily filter my incoming webmentions to view them based on who sent them
#
aaronpk I don't need multiple endpoints to do that
#
melvster rhiaro: do you know how a webmention would be written out as RDF?
#
rhiaro melvster: uh, I haven't thought about it
#
rhiaro It's just a source and a target
#
rhiaro Addressees, eg. for private webmentions, would be discoverable from the post itself
#
rhiaro Whereas in ActivityPump, the entire post is sent to someone's inbox, and would include addressees as metadata
#
melvster so [] a <???:WebMention> ; <urn:source> <source> ; <urn:target> <target> .
#
melvster call by value vs call by reference
#
melvster it would be helpful to have a webmention ontology
#
melvster maybe put it in AS2?
#
rhiaro The webmention is just a notification. I don't know if it needs to be an object in its own right. The source post contains a mention of the target post.
#
rhiaro AS2 has Mention, though I'm not sure of it's intended use
#
rhiaro The webmention is just alerting the server of the target post that the source post mentioned it
the_frey joined the channel
#
melvster rhiaro: if WebMention was added to AS2 we could then send it to SoLiD containers too ...
#
melvster ideally source and target too
#
rhiaro You wouldn't send the webmention to solid containers, you'd send the post
#
melvster why?
#
rhiaro The source post
#
rhiaro The act of sending the source post *is* the webmention
#
rhiaro Or you'd send the URI of the source post
#
rhiaro melvster: have you seen https://github.com/w3c-social/activitypump/blob/master/implementation.md ? AP is still ongoing work, but I feel it might align with an LDP backend
#
melvster rhiaro: why couldnt SoLiD use webmention in exactly the same way as indieweb?
#
melvster oh thanks!
#
melvster so whats an outbox then?
#
rhiaro an outbox is just a feed of content
#
rhiaro indieweb equivalent is just the homepage
#
rhiaro homepage marked up with microformats, that is
#
rhiaro in AP, the outbox endpoint is also what a client sends new content to. (ie. when you use a client to create new content for your own feed)
#
rhiaro The equivalent in indieweb being a micropub endpoint
#
melvster rhiaro: ok I think the only difference between a SoLiD webmention and an indieweb webmention is that SoLiD returns a 201 and indieweb returns a 202
#
melvster i wonder if we should maybe use a generic rel="inbox"
#
melvster if a webmention endpoint really is an inbox
#
ben_thatmustbeme very different technologies
#
melvster ben_thatmustbeme: what are very different technologies?
#
ben_thatmustbeme http://lists.w3.org/Archives/Public/public-socialweb/2015Jun/0050.html
#
ben_thatmustbeme webmention vs inbox
#
ben_thatmustbeme in indieweb/solid/pump
#
melvster ben_thatmustbeme: arg! the whole conversation started with me asking if indieweb was going to implement inboxes, and then the answer was that webmention was an inbox ...
#
ben_thatmustbeme well an archive, and thats not really an answer, i think the answer was some time ago.
#
melvster ben_thatmustbeme: i dont care about indieweb/solid/pump ... i only care about the WEB ... and how it can be social ...
#
melvster the brand isnt important, so long as things are interoperable
#
melvster and so long as things scale
#
rhiaro ben_thatmustbeme: what's your perspective on difference between an inbox endpoint and a webmention endpoint?
#
melvster they can be both
#
ben_thatmustbeme rhiaro: i am no where near as familiar as you would be, but from what i understand, an inbox gets sent activity data directly, json encoded
#
ben_thatmustbeme and webmention endpoint requires extra poll to verify data, doesn't trust the message
#
rhiaro An inbox endpoint as currently specified by AP expects the full post/activity(/whatever, all the content + metadata) whereas webmention just expects the URL. But webmention then goes to pull the post itself to verify. AP verfies the post based on what it's sent. So same principle, just different implementation
#
melvster call by value vs call by reference -- surely that's an implementation detail?
#
ben_thatmustbeme melvster: its unlikely you will interoperate with everyone right now. the goal is to create a standard for interoperation in the future. You don't care about the brand, good. those are the current groups that have things working in various amounts of cobbled together implementation. You say you want to implement user stories, but you have it the wrong way around, what is implmented right now is sort of the bootstrapping for the future, we
#
ben_thatmustbeme implment, find what works and what doesn't, then we spec so that future implementations don't have a bad spec. So what exactly do you want to implement? Do you want to interoperate with current groups? pick one to start learning how those people are doing it. Do you want to build something new and experiment, excellent, try something new entirely and make notes of how it goes
#
ben_thatmustbeme our work in implementing is what is informing our decisions in creating the spec/user stories which will then lead to interoperable implementations.
#
melvster i just want to make social a first class citizen of the web
#
melvster if so things are interoperable by definition
#
melvster then there's the patterns
#
melvster so you actually get useful things working
#
melvster with whoever wants to participate
#
KevinMarks we have useful things working; please participate
#
melvster ben_thatmustbeme: I think webmention should probably return a 201 because you're creating some (unnamed) data structure on the server, that it's going to be future processed really doesnt matter that much, EVERYTHING on the web is future processed
#
melvster but i dont think our code will fail on a 202
#
melvster maybe 201 is wrong also ... i dunno
#
aaronpk 201 would indicate the we mention was successful, which isn't true until it is verified
#
Loqi brain fart
#
aaronpk 202 is an acknowledgement of the request and saying it will be verified outside of the http request
#
melvster i guess with 201 you have to return the resource
#
melvster but webmentions are a bnode
#
melvster aaronpk: i thought indieweb likes to give things urls, why not give each webmention a url, and return it to the POSTer?
#
aaronpk May or may not be
#
aaronpk I do
#
rhiaro melvster, I treat webmentions differently depending on what kind of mention they are
#
aaronpk Go try t
#
rhiaro if I get a direct reply, I just add <target> sioc:has_reply <source> to my store
#
aaronpk You'll get back a url that you can view to check the status of the web mention request. It tell you if its processing or verified or failed and why
#
ben_thatmustbeme i have queue entries, so i return a URL to those, even though they don't last
#
ben_thatmustbeme but thats a queue, not a webmention. webmentions don't actually exist, its a post request
#
ben_thatmustbeme s/its a/its Just a/
#
melvster they do exist, they might not be long lasting
bblfish joined the channel
#
aaronpk melvster, please go implement webmentions before suggesting changes to the protocol
#
melvster aaronpk: im trying to work out how to implement inboxes, and you said webmention was an inbox, well turns out they are pretty different maybe
#
melvster aaronpk: isnt one of the goals of this group to standardize webmention?
#
KevinMarks the way to resolve that maybe is to implement webmentions and see how they map to your inbox data model
#
rhiaro 'an inbox' according to AP hasn't been implemented yet, so implementing webmention to see if is a sufficent inbox endpoint would probably be a good place to start
#
melvster KevinMarks: I just did and it returns a 201 ... so that's what we're discussing ...
#
melvster the thing is that SoLiD will always give you back a URL when you POST something
#
melvster webmention it seems aaronpk does but others do not
#
KevinMarks do webmention callers do much with the return value?
#
melvster i think webmention should be able to return EITHER a 202 (queued) OR a 201 (created) which sends back the URI of the created webmention ...
#
melvster we dont really have the concept of 202 because everything in the social web is in some kind of queue
#
aaronpk That is exactly what webmention says
#
aaronpk Depending on whether you verify a synchronously
#
melvster oh good
#
aaronpk It would be helpful if you read the spec
#
melvster well in that case we've implemented it already modulo the syntax
#
melvster aaronpk: if it was a standard id know where to look, i was reading : https://github.com/converspace/webmention/blob/master/README.md
shepazu joined the channel
#
melvster anyway thanks for the clarification
#
aaronpk I assume you meant "a w3c standard"
the_frey joined the channel
#
melvster just a stable uri like an rfc
#
melvster normative
#
aaronpk http://indiewebcamp.com/webmention works
#
melvster in that case solid has implemented a semantic webmenion already then ... maybe if we all start using JSON it all converges into one
#
aaronpk or http://webmention.net
#
melvster aaronpk: i was looking at http://webmention.org/ ... hard to know that was the WRONG url
#
melvster anyway thanks for the pointer
shepazu joined the channel
#
melvster so I would say an "inbox" could accept
#
melvster - a webmention
#
melvster - an activity stream object
#
melvster - a SoLiD object
#
melvster that would work
#
melvster aaronpk: does that work for you?
#
ben_thatmustbeme not for me
#
aaronpk re: an activity stream object, why not just send a URL pointing to the activity stream object?
#
melvster yeah you could do that too ...
#
ben_thatmustbeme we should work out something that works best and standardize on that
#
ben_thatmustbeme not create a spec that puts a lot of work on the implementer
#
aaronpk right. there needs to be an advantage to me to do anything different than I do now
#
aaronpk this is how twitter and facebook wedged their proprietary meta tags into tons of web pages, they gave them an incentive, they didn't just say "here is a great standard you should do it"
#
melvster but we've already decided to use activity streams, right?
#
melvster i thought that was consensus?
#
melvster so if activity streams can contain a webmention that kills two birds with one stone
#
melvster and we can create inboxes
#
melvster if we think a webmention endpoint is a kind of inbox this works well
#
melvster next question: how to find an inbox
tantek joined the channel
#
tantek good afternoon #social!
#
tantek reads today's logs
#
melvster https://www.youtube.com/watch?v=WQLzZf34FJ8
#
melvster worth watching ^^
#
melvster imho the state of the art in the social web
#
tantek nah, the state of the art in the social web would be a bunch of actual permalinks of posts by people, something readable / referenceable / indexable, not just a youtube video
#
melvster tantek: did you watch it?
KevinMarks joined the channel
#
tantek melvster: nah - can't be bothered with a video that starts with "let me start with the story of, in 2011… " (yawn)
#
melvster he talks about how they started with a RESTful API ... then how they needed to go beyond that
#
tantek I have a better such story: In 2011, we started IndieWebCamp and getting people to actually own their own content on their domains, and interoperate.
#
tantek no PowerPoint needed
#
tantek melvster - RESTful meh http://indiewebcamp.com/RESTful#Criticism
#
melvster 260 billion requests per day :)
#
tantek I'm sure MySpace used to have billions of requests for day too. Meh.
#
tantek melvster: you said "only care about the WEB ... and how it can be social …" http://socialwg.indiewebcamp.com/irc/social/2015-07-06/line/1436213594650 - in that case, could you add yourself with link to your *social* *WEB* site to https://www.w3.org/wiki/Irc-people ?
#
melvster tantek: done
#
tantek the link looks good - I think the trailing "|" is creating a broken image however, if you want no image perhaps leave it out, or re-use the image URL you have embedded on your home page
#
tantek melvster re: "we've already decided to use activity streams, right? / i thought that was consensus?" http://socialwg.indiewebcamp.com/irc/social/2015-07-06#t1436216690289 - not quite, we have consensus on *publishing* AS as a W3C draft, which means we have consensus to publicly discuss it as a work item of the Social Web WG.
#
tantek there's certainly no consensus on actual *use* of AS, no actual W3C AS2 consuming code yet (except maybe the test validator thing jasnell built).
#
tantek "consensus on use of (insert technology x)" is determined by people shipping things that use technology x
#
tantek s/shipping/deploying live somewhere on the web we can see it and check it etc.
#
melvster tantek: but the consensus is that the social api will use activity streams, right?
#
tantek there's no such consensus AFAIK
#
tantek and given the lack of implementer interest (like maybe 1?) in AS2, but in contrast, the multiple implementations of Social API candidates (e.g. > 10 implementations of Micropub), it is more likely that AS would be remodeled based on Social API
#
melvster tantek: ok so what about JSON ... it says in the charter JSON will be used, is that consensus yet?
#
tantek charter says at least JSON will be used for a social syntax - and various social syntax alternatives use various JSON variants or have canonical JSON representations
#
tantek JSON is kind of like a lowest common denominator interop touchpoint of the current decade. Like XML was last decade, and something else will likely be next decade.
#
melvster tantek: so everyone will support JSON, right?
#
tantek hard to predict, depends on the timescales. Can only answer today based on current deployments. E.g. every interop IndieWeb site publishes microformats which are available as canonical JSON from a plethora of open source parsers.
#
tantek so if that qualifies as "support" for what you're looking for, then yes in that way, for current sites using those techniques.
#
melvster i think support means consume and produce, so I think that's currently a *no* for indieweb today, but maybe in future ... actually SoLiD is in the same boat ... but we're working on a JSON LD parser
#
melvster tantek: thanks for clarification
#
tantek more importantly, it's our job as a W3C working group to help standardize what appears to be working, rather than force some topdown ideological point of view
#
tantek so for example, if tons of sites have consuming and producing HTML+microformats working interoperably, then it really doesn't matter that the charter asserts a topdown requirement of JSON for a social syntax
#
tantek running code trumps any committee charter
#
tantek that just means the charter needs an update to reflect current / growing market realities
#
tantek in the interim, we can help specify conversion, e.g. using canonical microformats JSON
#
tantek when there are various interop SoLiD sites with different implementations, then that too is something the WG should look at to see what the WG can help standardize
#
tantek but first things first, for now the only charter update we're working on is to use the new W3C Document & Software license for all our specs. hoping sandro will have an update for us on that tomorrow. speaking of I should put that on the agenda
#
melvster tantek: well let us know if you anticipate the charter changing from what it is today
#
tantek melvster: I do, to use the new license
#
tantek per group resolution last week and before
#
tantek assuming AC approval, which should go smoothly given that the AC approved the use of the new license in general for W3C work
#
melvster tantek: yes you said, i mean other aspects, such as using JSON ... im not sure 10 users, even as if they all implement micropub, is a convincing reason to change the charter
#
aaronpk not 10 users, 10 different implementations
#
aaronpk at least one of which is deployed to thousands of users :)
#
melvster which?
#
aaronpk withknown.com
#
Loqi Tantekelik made 1 edit to [[Socialwg/2015-07-07]] https://www.w3.org/wiki/index.php?diff=84955&oldid=84931
#
aaronpk the wordpress plugin is doing well too https://wordpress.org/plugins/micropub/
#
melvster aaronpk: can you point me to a profile?
#
tantek melvster, it's likely in the short term that even if we relax the JSON requirement, it will stay as some sort of "should" (e.g. should have a canonical representation in JSON) given the the current fashionability of JSON in snowflake APIs.
#
aaronpk a profile?
#
melvster one of the 1000s of users
#
aaronpk http://werd.io/ :)
#
tantek my understanding is that all the various social syntax / social API approaches do have some sort of conversion to / canonical version of whatever they produce/consume in a lowest common denominator JSON form
#
tantek perhaps not published directly, but available through defined specs / open conversion proxies / parsers etc.
#
tantek e.g. here's some "published" JSON if you like: http://pin13.net/mf2/?url=http://werd.io/ (building upon what aaronpk just gave you)
#
melvster tantek: that is a bit of a mess ... what does "me" actually mean ... he has "me" and an email address in there ...
#
melvster tantek: second question can this JSON handle numbers and strings, or just strings?
#
tantek sure, the "me" relation is used to reference other URLs that also represent the same person. see also: http://microformats.org/wiki/rel-me
#
tantek in JSON numbers work best as strings
#
melvster just FYI: a URL doesnt represent a person, it denotes a person, it's a name, not content
#
tantek melvster: do you have some sociological data to backup that assertion? my own anecdotal experience is otherwise, that for over a decade now, bloggers etc. use URLs to represent themselves
#
tantek and of course current indieweb usage as well
#
melvster im thinking or represent as in representaion as in REST ...
#
tantek REST is meh per previous citation
#
melvster ill defer to bigbluehat on that one :)
#
melvster REST is the architecture of the web, it's what lets it scale
#
melvster tantek: im inferring from your answer that the JSON is string only, and not number and strings
#
tantek melvster - seems reasonable. I have seen snowflake APIs that use JSON (e.g. Twitter) switch from use of numbers to numbers-as-strings over time - so that makes sense
#
melvster i believe AS2 can handle numbers too
#
melvster that's great for twitter, but other groups at the w3c e.g. payment, security, credentials I believe have use cases that use numbers
#
tantek melvster - I think you need a citation for claiming "REST is the architecture of the web…" - as most of the web doesn't use REST (trivial to test by trying various REST verbs at a random sampling of URLs)
#
melvster so i wanted to check before hand if that work easily converts
#
tantek more like, REST is ignorable
#
tantek as in may ignore, not MUST ignore
#
tantek the web has scaled just fine without depending on REST
#
tantek such number use-cases in practice over time result in being re-purposed as strings - that was my point about Twitter
#
tantek it's great that groups think they have aspirational needs for "number" type, yet in practice, any such use of JSON eventually gets replaced by number-as-string
#
melvster i use number in deployed code today
#
melvster any payments system pretty much will need numbers
#
tantek sure, wasn't disputing that there's numbers in deployed code. my point was about eventual evolution to strings
#
tantek melvster: nah, especially in payment systems, due to floating point rounding errors etc. strings and special currency processing systems end up being necessary for actual financial-defined precision
KevinMarks joined the channel
#
melvster huh?
#
melvster $99.99
#
melvster 2 decimal places is fine for most needs
#
melvster or integers
#
melvster tantek: i think the JSON will be converted via mime types etc. ... so long as indie web can be converted to json by the consuming client and is proven to be interoperable I think it's going to be OK
#
melvster i think it's going to be important to agree on a JSON based social syntax tho
#
melvster and make sure the convertors give the right thing
jasnell, melvster, jasnell_ and bblfish joined the channel