#social 2015-02-25

2015-02-25 UTC
#
Loqi Abasset made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82392&oldid=82390
#
Loqi Abasset made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82394&oldid=82392
caseorganic, KevinMarks and tilgovi joined the channel
#
Loqi Benthatmustbeme made 2 edits to [[Socialwg/Social API/More user stories]] https://www.w3.org/wiki/index.php?diff=82396&oldid=82384
#
Loqi Benthatmustbeme made 2 edits to [[Socialwg/Social API/More user stories]] https://www.w3.org/wiki/index.php?diff=82398&oldid=82396
caseorganic and bblfish joined the channel
#
Loqi Benthatmustbeme made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82400&oldid=82394
#
Loqi Benthatmustbeme made 1 edit to [[Socialwg/Social API/More user stories]] https://www.w3.org/wiki/index.php?diff=82399&oldid=82398
KevinMarks joined the channel
#
ben_thatmustbeme Only 2 hours left to get your votes in!
AnnB joined the channel
#
AnnB here we go .. down the home stretch
#
Loqi Tantekelik made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82402&oldid=82400
harry joined the channel
#
aaronpk tick tick tick
#
Loqi Aaronpk made 5 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82407&oldid=82402
#
Loqi Abasset made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82408&oldid=82407
bblfish joined the channel
#
Loqi Abasset made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82410&oldid=82409
#
Loqi Tantekelik made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82411&oldid=82410
#
Loqi Aaronpk made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82409&oldid=82408
#
Loqi Tantekelik made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82412&oldid=82411
#
Loqi Aaronpk made 4 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82418&oldid=82413
#
Loqi Abasset made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82416&oldid=82412
#
Loqi Aaronpk made 4 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82424&oldid=82419
#
Loqi Abasset made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82421&oldid=82418
tantek joined the channel
#
aaronpk alright that's all i've got in me
#
tantek I think I may need to do a blog post on how "social web" doesn't just mean everything on the web
#
Loqi Abasset made 3 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82429&oldid=82425
#
Loqi Aaronpk made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82427&oldid=82424
AdamB joined the channel
#
ben_thatmustbeme AnnB. On mobile so it's easier for me to respond here. Re: https://www.w3.org/wiki/Socialwg/Social_API/User_stories#Contact_Info I actually do
#
ben_thatmustbeme ben.thatmustbe.me/contact has a lot of info that is only available to some people after they authenticate to my site
#
Loqi Abasset made 3 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82432&oldid=82429
#
ben_thatmustbeme Tantek and aaronpk could auth and then get my cell number for example
#
AnnB hiya ... would rather not chat right now .. trying hard to get our votes in!
#
ben_thatmustbeme Okay. Understandable
#
Loqi Abasset made 1 edit to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82433&oldid=82432
#
AnnB FYI .. the aboyet comments on the last ones are in collaboration with me...
#
Loqi Aboyet made 4 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82437&oldid=82433
bblfish joined the channel
#
AnnB 2 more to go .... tick tock tick tock
#
Loqi Aboyet made 3 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82440&oldid=82437
#
AnnB yayyyy we did it!
#
Loqi woot
#
Loqi Aboyet made 3 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82443&oldid=82440
#
AnnB 25 seconds remaining
#
AnnB haha
#
aaronpk wow nicely done
bblfish joined the channel
#
AnnB big pats on backs to all who voted!
#
AnnB g'night .. (Seattle time)
tantek, bblfish, jaywink, the_frey, pfefferle, KevinMarks, timbl and barnabywalters joined the channel
#
melvster "everyone uses Slack, HopChat, Gitter, you name it! we also use IRC in this very group... — Pavlik elf (talk) 08:46, 23 February 2015 (UTC)" +1
bblfish, pfefferle and almereyda joined the channel
#
Loqi Eprodrom made 2 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82448&oldid=82443
#
ben_thatmustbeme melvster, many use git too, this very group does. doesn't mean the social API should handle code bases
#
ben_thatmustbeme i would love a decentralized chat, don't get me wrong, but it shouldn't be a priority of v1
#
melvster ben_thatmustbeme: ive already coded decentralized chat ... was just using it
#
melvster just saying that 3-5 years ago chat was state of the art, now it's an expectation of social systems, so there's a gap there between what this group is doing and what people are using today, but i understand only so much can be implemented
#
melvster doesnt matter too much it's all going to get implemented in parallel, then there can be a v2.0 spec maybe
#
melvster imho, realtime stuff just really adds a nice tight feedback loop ive found that people like it
#
ben_thatmustbeme melvster, what do you use for decentralized chat? Open Source? i'd love to check it out
#
melvster ben_thatmustbeme: I use bblfish's webid and Arnaud's linked data platform (GOLD implementation), im currently testing it but am pretty happy with it so far ... if you get yourself a webid and upload an avatar, id be happy to test with you
#
melvster yes it's all open source MIT license, tho the latest version isnt quite checked in yet, i could do so if anyone wanted to hack on it
#
melvster still bug crushing :)
#
melvster the realtime updates work via a websocket
#
melvster ben_thatmustbeme: do you have an online profile?
#
ben_thatmustbeme melvster, i have ben.thatmustbe.me, everything is mf2
#
ben_thatmustbeme i'm wondering how this could integrate it
#
melvster nice page
#
ben_thatmustbeme i could generate a foaf file and link via rel="me" but it seems to be duplicating a lot
#
melvster is there any w3c linked data in there, what I need is a Person, a name and an avatar, right now
#
melvster tho soon ill be adding default avatars for those that dont yet have one
#
ben_thatmustbeme well, the css on it is a mess, but I haven't really had that much time to play with that, i'm always more interested in the backend things
#
melvster or you could do content negotiation
#
ben_thatmustbeme not as LD, just mf2
#
ben_thatmustbeme that too
#
melvster i dont support that right now, maybe one day tho
#
melvster id recommend thinking about having a foaf file to augment your profile, but if you want one to play around with, i can suggest here: http://cimba.co/
#
ben_thatmustbeme where is the github, i'd love to add it, much of the foaf file is overlapping with an hcard, just requires a minimum of certain things
#
melvster the sign up tab will generate one for you and give you an access controlled linked data platform storage area
#
ben_thatmustbeme i'm on the foaf-a-matic to see what it thinks of as "me"
#
ben_thatmustbeme actually it would be cool to create a service to generate foaf from mf2, mf2tofoaf.com/ben.thatmustbe.me
#
ben_thatmustbeme something like that
#
melvster ah nice let me show you an example
#
melvster this is a pretty good example of FOAF imho : https://deiu.rww.io/profile/card
#
ben_thatmustbeme oh, there is a rel=meta for it already
#
rhiaro haven't read context from last few minutes of conversation yet, but thought I'd jump in with my foaf profile (not that there's a lot on it): http://rhiaro.co.uk/about/#me My conneg is broken for some reason, so see also http://rhiaro.co.uk/about/me.ttl and http://rhiaro.co.uk/about/me.html
#
ben_thatmustbeme rhiaro, looking at foaf profile from indieweb perspective and how to roll the ideas in together. perhaps generate foaf from mf2 to allow interop
#
rhiaro yeah, I'm extremely interested in interop between indiweb and linked data
#
ben_thatmustbeme violates DRY if its not automatic
#
ben_thatmustbeme generating it that is
#
rhiaro generating from mf2 sounds good
#
rhiaro I'm linked data before indieweb, so prefer foaf to mf2, but that's just me
#
rhiaro I also prefer linked data vocabs for the extensibility/self-documentation/discoverability; coming from that, mf2 seems preeetttty centralised to me
#
ben_thatmustbeme mf2 uses the css model of anything not standard is just prefixed for extensibility
#
ben_thatmustbeme so p-media is a standard, x-myapp-xyzfile is extension
#
Loqi Akuckart made 10 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82462&oldid=82448
#
rhiaro Okay, fixed my conneg
#
ben_thatmustbeme aaronpk and I have started doing the same with mobilepub as we are finding clients are adding new features that don't make sense to centralize
#
rhiaro Yes, but there's no way of discovering what something with a prefix means or is for
#
rhiaro If I see class="x-floop" either by inspecting someone's code manually or by parsing it with a script, I have no way of finding out what a floop is without asking the person
#
rhiaro but namespacing with URIs, assuming they're used properly to link to documentation of terms, resolve this
#
rhiaro both for humans and scripts
pfefferle joined the channel
#
ben_thatmustbeme if you see x-floop you shouldn't have to look it up at all, you have some object floop already parsed, you see what data they store in a floop. why do you need a definition beyond that?
#
rhiaro Hm, maybe I need some concrete examples of extensions
#
rhiaro But with additional docs someone could provide guidence on how best to display the content perhaps
#
ben_thatmustbeme i haven't needed them personally yet. actually I hit that with x-audio and x-video which were not standardized at the time
#
ben_thatmustbeme display is UI, thats on you. mf2 is how to get the data into a machine readable format
#
Loqi Akuckart made 8 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82471&oldid=82462
#
rhiaro but then when someone has it in machine readable format, they might want to display it again somewhere, like in a reader
#
ben_thatmustbeme well either your reader supports floops across the board or it doesn't, you see a floop, you mark it up once, you are done
#
ben_thatmustbeme i'm trying to think of a good example
#
ben_thatmustbeme i think any of the extended post types is where this would really come in to play
#
ben_thatmustbeme x-badge maybe
#
ben_thatmustbeme its still going to have content, that is text to display it, and maybe an image to show, those are going to be easy enough to display
#
rhiaro maybe I'm over complicating it for indieweb use cases. I like the idea of a reader that could display any content, any post type, without prior knowledge, if it was documented appropriately by the publisher
#
ben_thatmustbeme it should still fall back to just displaying the content without having to do custom mark-up
#
rhiaro I agree with fall back
#
rhiaro s
#
ben_thatmustbeme but a data source deciding the display on my service would annoy me
#
rhiaro I can see that
#
ben_thatmustbeme have a link to a definition would only really get you the set of maybe required fields
#
ben_thatmustbeme but it won't take long to find someone who just ignores required fields and then you are stuck how to display that
#
ben_thatmustbeme assume the markup could always be dirty, because it will be
AnnB joined the channel
#
rhiaro that's why fallbacks. But there could be enhancements for clean markup
#
ben_thatmustbeme not sure I follow that
#
rhiaro Maybe if you had a custom reader that only displayed metric posts gathered by hardware, and the reader could see that x-weight is input by smartscales from the documentation, and knows to include it. But maybe this goes into classes/rdfs world out of scope for indieweb.
#
Loqi Akuckart made 7 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82478&oldid=82471
#
rhiaro sorry that was a new vague use case, not an expansion on what I previously said
#
ben_thatmustbeme hmm, actually looking up how aaronpk does his weight markup
#
rhiaro I just checked that, looks like p-weight made it into mf2
#
barnabywalters rhiaro: I am intrigued by the idea that linked data would allow (for example) “readers to display post types they’re not programmed to by fetching the discoverable definition”
#
rhiaro and p-bodyfat, which seems pretty obscure
#
barnabywalters how does that actually work? is there an example somewhere?
#
rhiaro Uh, not off the top of my head. I should try to implement something.
#
ben_thatmustbeme yeah, i'm checking that now, he might have just not written is as an extension, tsk tsk
#
barnabywalters I’d love to see some documentation telling format-creators how to tell consumers how to display something, and vice versa
#
rhiaro http://microformats.org/wiki/measure-brainstorming
#
AnnB I'm loving this convo .. you guys are great
#
AnnB "p-bodyfat" ... sheesh ... must be from the runner in the crowd
#
barnabywalters (not least as I’ve heard this use case as a reason for linked data several times but have never seen anyone able to demonstrate any meaningful implementation)
#
AnnB I seriously need a demo of the best collab tools outside the enterprise
#
rhiaro this is where I find meeting of indieweb and linked data extremely interesting, because indieweb has lots of webby implementations and linked data has lots of terrible looking academic things
#
ben_thatmustbeme i figured weight would be one, but i think that should be x-bodyfat
#
rhiaro I'm dreaming of mashing them together
#
ben_thatmustbeme i can't see that being something very commonly used
#
barnabywalters I’d love to see even a terrible looking academic demo :)
#
barnabywalters the few times I’ve seen people actually demo things built on linked data they’ve usually been quite impressive
#
rhiaro I can't think of anything that say, provides a UI template in the term definition, but there *must* be something
#
rhiaro goes searching
#
barnabywalters fwiw, the microformats, HTML-first approach solves this use case by using HTML or plain text fallback content and progressively enhancing from there
#
barnabywalters “likes” are a good, widely implemented example: consumers who understand them present them as a “like”, but consumers who don’t simply see them as a text comment saying “X liked this”
#
rhiaro Yeah, that's an approach that makes all the sense
#
rhiaro So the next level of enhancement would be to say "a floop is a like with color: green" and a reader that could understand that could act accordingly
#
rhiaro need better than silly examples though..
#
Loqi Akuckart made 9 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82490&oldid=82478
#
barnabywalters rhiaro: agreed, always best to build off real-world use cases :)
#
rhiaro I wonder where elf-pavlik is, he probably has some ideas..
#
rhiaro !tell elf-pavlik You use rdfa and mf2 in your site, what are the specific practical benefits of each for you?
#
Loqi Ok, I'll tell them that when I see them next
#
rhiaro !tell elf-pavlik Help brainstorm about linked data/indieweb overlap :)
#
Loqi Ok, I'll tell them that when I see them next
#
barnabywalters tommorris is another user of both rdfa and mf2, on https://tommorris.org/
#
rhiaro https://docs.google.com/viewer?url=http://ceur-ws.org/Vol-1026/paper7.pdf
#
rhiaro "Different resources need different UIs, e.g. a person could be displayed with a profile page, whereas a place should be displayed with a map."
#
rhiaro Disclaimer: haven't actually read it all yet
#
rhiaro Oh also bbc.co.uk/things displays places with maps and other things without
#
ben_thatmustbeme a place could be displayed as a profile page (their actual website), map, a photo, text. a person could be displayed as a profile page, a point on a map, a photo, text. Certainly different things need different UIs, but given lack of knowledge of a item you don't trust UI to someone else
#
Loqi Akuckart made 6 edits to [[Socialwg/Social API/User stories]] https://www.w3.org/wiki/index.php?diff=82496&oldid=82490
#
barnabywalters only thing I can find in that paper which might be an implementation of what we’re discussing is http://www.w3.org/2005/04/fresnel-info/
#
rhiaro I can see that most people implementing readers would want control over what their content looked like, but I can also see that the publisher of some data might have a better idea over what something should look like
#
rhiaro One of my friends is turning all of wikihow into linked data for his phd. Which is a good link between normal-people-web and linked data. Maybe I can pull a use case out of that.
#
AnnB wow, that'd really be cool
Augier joined the channel
#
rhiaro AnnB: it is cool. In fact, he's already done it. Now he's trying to figure out how to make it academic :) Looking at automatic task composition etc
#
rhiaro his common example is how to make a pancake, and our favourite use case for pancake making steps as linked data is that a robot can understand it and do it
#
AnnB hmm ... academic ... will be interested to chat about that in Boston
#
rhiaro one day!
#
AnnB it might be easier to instruct robots than humans, in a sense
#
AnnB they only do what you tell them to do
#
AnnB humans .. who knows
#
rhiaro A robot could start with no info, get the steps as linked data, and from that understand eg. what an ingrediant is, then look up flour and find out where to get it, etc
#
rhiaro all we need is the hardware really :D
#
AnnB those cooking robots ... can't quite pick them up at the store yet
#
rhiaro http://www.eatmedaily.com/wordpress/wp-content/uploads/2009/06/robo-chef-pancake.jpg
#
melvster reading back
#
melvster ben_thatmustbeme: my indeiweb profile is also my FOAF http://melvincarvalho.com/ so i can use indie auth etc. but also have foaf in the html via RDFa and content negotiation to give turtle so that I can use all the web 3.0 stuff
#
melvster i dont use the rel="me" pattern
#
melvster im not sure how much tooling supports that
#
melvster rhiaro: this is the viewer i tend to use for linked data: http://graphite.ecs.soton.ac.uk/browser/?uri=http://rhiaro.co.uk/about/me.html#me
#
rhiaro better http://graphite.ecs.soton.ac.uk/browser/?uri=http://rhiaro.co.uk/about#me
#
rhiaro It didn't occur to me graphite did customised views based on types
#
rhiaro I thought it just pretty printed text
#
melvster very nice
#
melvster rhiaro: could I add your link to my homepage?
#
rhiaro sure! linking++
#
melvster great let me do that
#
rhiaro I'm going to add more to my foaf at some point
#
melvster very nice foaf
#
melvster you even have a fragment id in there -- bonus points!
#
melvster done
#
rhiaro I had a webid cert in there but then I forgot to back it up and formatted my laptop, so I need to generate that again. I also had some general problems with myprofile.eu
#
rhiaro I forget what precisely, this was ages ago.
#
melvster there's a new system now called rww.io
#
melvster and one at databox.me
#
melvster you can replace certs, someone is working on recovery via email which will be quite nice
#
rhiaro I was on rww.io too. I really wanted to get the linked data calendar working, but for some reason could never quite get everything to gel
#
melvster i like the calendar
#
melvster rww.io is a great place to keep your profile, someone is working on a profile editor too
#
rhiaro So I can't get rhiaro.rww.io because I already have that but lost the cert - can I regenerate the cert, or is it gone forever? (The auth stuff is a bit beyond me)
#
melvster yes it can be fiddly to get things working, as the code is quite new, and the team is not the biggest, but hopefully little things get fixed more and more
#
melvster i am hoping 2015 will be a break out year ... we will see
#
rhiaro And, when I try to sign up it says my webid is rww.io/profile/card#me, but I want rhiaro.co.uk/about#me as my webid... I think that was what I couldn't figure out before
#
rhiaro Or am I just supposed to sameAs them and deal with multiple URIs?
tantek joined the channel
#
rhiaro Anyway, gotta run, time to teach kids to code at prewired.org :)
#
melvster ok answering questions in order ...
#
melvster oh ok ... you can host your profile where you want
#
rhiaro I'm always signed in to irc, so leave me messages and I'll be happy to chat when I'm not elsewhere!
#
melvster then you put the public key in your profile, where you want your identity
#
melvster ok great, chat soon! :)
Augier joined the channel
#
melvster from my experience indieweb and linked data do seem to work pretty well together, if you want the indie web functions add the indieweb markup, and if you want the linked data functions add that markup, i do both, and ive not had a conflict yet
#
melvster re robots we were trying to train quad copters to fetch coffee :)
almereyda, timbl_ and caseorganic joined the channel
#
aaronpk good morning
caseorganic joined the channel
#
tantek good morning aaronpk
#
aaronpk did everyone get their votes in in time? :)
#
AnnB in the nick
#
AnnB now will be very interesting to study what we all put
#
ben_thatmustbeme I tried to add in some alternate versions to things as well
#
ben_thatmustbeme give the More User Stories page some love
#
AnnB has to go do her 'day job' now ... keen to read this chat later
#
tantek got all my votes in - may finesse some of the explanations
#
tantek which I think is fine
caseorganic and the_frey joined the channel
#
melvster tantek: you dont think chat is part of the social web these days?
#
tantek melvster: "chat" is too broad
#
tantek and could mean lots of things
#
melvster instant message?
#
tantek so no, in the most liberal interpretation
#
tantek direct message / instant message appears to be fairly common in social networks
#
tantek however does that mean it is social web?
#
tantek it's not "on" the web per se
cwebber2 joined the channel
#
tantek messages in direct message / instant message don't have permalinks you can reference *on the web*
#
tantek so no, chat is not web
#
tantek is that not obvious?
#
tantek web is HTML delivered over HTTP at a URL
#
tantek if something lacks any of those, it's not web
#
tantek HTTP(S) that is, obv
#
melvster so if instant message had perma links, it would be part of the social web?
#
tantek melvster - would be closer to, yes
#
tantek but no one's implemented them that way
#
melvster tantek: did you see elf's comment?
#
tantek so no
#
tantek I'm a bit surprised that so many people think "social web" means "anything you do with a web browser"
#
melvster "everyone uses Slack, HopChat, Gitter, you name it! "
#
tantek so much scope creep in the user-stories
#
tantek melvster, not everything "social" is social web, and not everything "web" is social web.
#
tantek I would think that would be obvious
#
tantek or do so many people not care about scope creep?
#
melvster i dont know about hopchat, but slack and gitter use perma links
#
melvster the chat app im writing also uses perma links
#
melvster it would not surprise me if facebook chat also had perma links
#
tantek melvster: AFAIK it does not
#
tantek so surmising is not helpful
#
tantek also - whether something *internally* uses permalinks is very different than whether it exposes them in the UI
#
tantek internal stuff is just plumbing, and can change at will
#
melvster but gitter and slack do, as does my app
#
tantek whereas something on the web has a URL as part of its UI
#
melvster i dont know how much you use facebook, but there is or used to be a message view, where you can see historical messages
#
melvster anyway, unsure on that one
#
tantek right, unsure means punt
#
tantek only act on things for which there is solid evidence
#
melvster "but no one's implemented them that way" <--
#
melvster let's leave facebook aside for the moment
#
melvster let me reprhase the question
#
melvster if chat / instant message had perma links, would you consider them part of the social web?
tilgovi joined the channel
#
tantek melvster, see above. necessary but insufficient.
#
melvster tantek: thanks for sharing your view, personally a few years ago i would probably have been unsure, but I agree with elf, that these days it's becoming more and more of an expectation, i understand of course not everything can be implemented
#
tantek nor should "everything" go into a v1!
#
tantek I think that's the larger point a lot of people are missing
#
tantek it's more important to ship a tight small v1 core ASAP
#
tantek and then iterate
#
tantek than it is to try to cover "everything that exists even currently"
#
tantek also, smaller API = smaller attack surface = more secure
#
tantek this is standard agile / iterative development methodology
#
tantek also, if we can modularize this API, even better
#
tantek I believe this disconnect is due to most people in the WG not being actual implementers of such APIs (server) or implementers of client web sites / web apps that *use* such APIs
#
tantek and thus there's a lot of "but I want this to exist!!!" going on, rather than "I plan to implement this ASAP"
#
tantek very different (and opposing) perspectives
#
melvster well ive already implemented chat, im in the testing phase now
#
tantek and frankly, to actually get a REC, one of those perspectives is MUCH more important than the other
#
tantek good!
#
tantek that's a strong vote then
#
melvster turns out testing in a distributed environment gets much more difficult
#
melvster i get different error cases for each user, but hopefully fixing them one by one
#
melvster tantek: thanks for clearing that up, I also agree only to put in scope what it thought to be achievable and implementable by the group, was just thrown a bit by your comment "-1 this is IRC, not social web. Tantek Çelik" -- and wanted to understand what you meant by that
#
tantek "testing in a distributed environment gets much more difficult" that's for sure! with pretty much anything!
#
tantek ah - chat *groups* is even more out of scope IMO
#
tantek for a v1 at least
#
ben_thatmustbeme a better question might be, should chat / instant messaging be in a different, more real time API
#
tantek e.g. I know FB has a notion of creating instant messaging groups
#
tantek unnamed even
#
tantek but e.g. Twitter does not
#
tantek or maybe they are talking about doing it soon?
#
melvster in my case the chat is agnostic to who uses it ... ive implemented a 1 person chat, which I use as my personal diary, a 2 person chat which I use for instant message, and a multi user chat which could be like a chat room / irc / gitter / slack ... the app doesnt really care too much who uses it
#
melvster in fact the chat messages are all just blog posts
#
melvster or the same format as a blog post
#
melvster with a slight subclassing
#
melvster so the same tech can be used for publishing
#
melvster but i found adding realtime via a websocket has been a great user experience so far
#
melvster ben_thatmustbeme : id love to see a social web API for realtime, not XMPP, but web based of HTTP and ws
#
melvster i guess it was tried with ostatus / pubsubhubub, but that kind of seemed to fail
#
rhiaro Just dropping in to note my comment on Chat rooms user story "Is this just posts and replies, with a UI (ie app-specific function) that makes it instant?" :)
#
rhiaro by the way melvster if you need users of distributed chat to help test, I'd be happy to
#
tantek melvster: I do think "real time" has additional constraints / demands that other aspects of the API do not have.
#
melvster great!
#
tantek thus it may better serve the use case to make it a modular addition to a core API
#
melvster rhiaro: I found a slight issue with your profile ... there's two different URLs and it's confusing my app
#
tantek rather than trying to patch posting/replying to be realtime
#
tantek and demand that everyone support realtime all the things
#
melvster rhiaro: one is http://rhiaro.co.uk/about/me.html#me and the other is http://rhiaro.co.uk/about/me.ttl
#
tantek melvster: so far pubsubhubbub has been quite difficult to consume
#
tantek based on implementer feedback
#
melvster tantek: agree!
#
tantek and the very few number of implementers, implementations
#
tantek I'm still unsure about PuSH
#
tantek but I do support the *publish* side of it from my site
#
tantek FWIW
#
tantek I don't know who if anyone is consuming that besides Superfeedr which feeds on all the things.
#
rhiaro melvster: so http://rhiaro.co.uk/about#me is my URI, and it is connegged to the html or ttl depending on accept header
#
rhiaro what does your app need in particular?
#
melvster tantek: what I am doing currently is that social web spaces have a websocket link relation "updates-via" and then you connect to that and tell it which pages you are interested in using 'sub <URI>' ... when that page changes you get 'pub <URI>' over the socket ... it's basic but is working for me
#
tantek melvster: aaronpk has some experience with implementing real time comments on posts
#
tantek I am not nearly as knowledgable in practice about it
#
melvster ah ok thx
#
tantek you may have better luck discussing that with him
#
tantek though I'll happily listen and learn
#
melvster rhiaro: the thing is that when I dereference http://rhiaro.co.uk/about/me.ttl it has data about the other page, and when I try curl on the other page with an accept header it doesnt give me back the turtle
#
melvster if i could do curl -H "Accept: text/turtle" http://rhiaro.co.uk/about/me.html and it gave me what's in the .ttl file that would be perfect
#
melvster maybe some .htaccess modrewrite such as RewriteCond %{HTTP_ACCEPT} ^.*text/turtle.*
#
melvster but i could fix my app to handle it I guess ... just havent considered that flow
#
rhiaro hang on, I think I might have an error in my .htaccess
#
rhiaro I broke it last week and hastily fixed it a couple of hours ago
#
rhiaro may have made a mistake
#
melvster ah
#
rhiaro But the idea was that curl -L -H "Accept: text/turtle" http://rhiaro.co.uk/about/ returns the me.ttl file
#
rhiaro hmm, weird, it looks like it should be working
#
melvster add on the me.html perhaps if that's where you want your identity to live
#
rhiaro It's a trailing slash thing
#
rhiaro curl -L -H "Accept: text/turtle" http://rhiaro.co.uk/about works
#
rhiaro I'll fix it for http://rhiaro.co.uk/about/
#
rhiaro Fixed!
#
melvster that was quick!
#
rhiaro So is http://rhiaro.co.uk/about#me okay for your app?
#
melvster ill test
#
melvster are you using a / or not?
#
rhiaro I wanted to have nothing to do with file extensions
#
rhiaro both should be fine
#
rhiaro I may be inconsistent when spreading the URI around... I should pick one
#
rhiaro I'll see what I have on bbc/things..
#
melvster about#me looks pretty good imho
#
rhiaro I'll stick with that then!
#
melvster ah so you're doing a 303 redirect
#
rhiaro yes
#
melvster works in tabulator
#
melvster im not sure if my library rdflib.js is picking it up tho, will test
#
melvster XMLHttpRequest cannot load http://rhiaro.co.uk/about. No 'Access-Control-Allow-Origin' header is present on the requested resource.
#
melvster rhiaro: any chance you could enable CORS?
#
melvster Access-Control-Allow-Origin : *
jaywink joined the channel
#
melvster tried it thru a proxy too and got 503 service unavailable
#
melvster tabulator can do it: https://deiu.rww.io/profile/card#me if you click in the arrow next to my url, then on 'about' ... you'll see your profile
AnnB joined the channel
#
rhiaro will do, brb in a bit
#
ben_thatmustbeme reading back a little of that conversation, we can currently do something of instant messaging with webmention + notification. as long as the webmention is sent right away. Look at my recursive context history posts. slap those in to private posts and you are close to done
#
ben_thatmustbeme https://ben.thatmustbe.me/note/2015/2/17/1/
#
ben_thatmustbeme the only issue is joining the conversation and not having to reply to specific posts is not really worked out
#
ben_thatmustbeme I also have webmention sends on a cron job rather than done automatically
#
ben_thatmustbeme hmm, seeAlso for friends from foaf would be a great place to find potential sites for webmention+vouch
#
ben_thatmustbeme i just whitelist and publish my whitelist with rel=me
caseorganic, ben_thatmustbeme and pfefferle joined the channel
#
rhiaro melvster: back in my office now. Enabled CORS in the htaccess file (shared hosting), should work now hopefully
#
melvster oh excellent!
#
rhiaro ben_thatmustbeme: I was chatting about vouch with someone a while ago. We didn't manage to figure out: if you're rejected from someone because they want a vouch, how do you know who to send as a vouch if they don't publish a list? Because as far as we can tell it's not required for someone to publish their list of trusted people
#
rhiaro Maybe that's a question to take to #indiewebcamp, but my instinct was to have a foaf:knows list anyway
#
rhiaro Of course rel= is probably the indieweb way
#
melvster rhiaro: sent you a PM
#
ben_thatmustbeme rhiaro, its a bit of discovery. I recommend posting friends lists, so you can find that easily
#
tantek rhiaro: more than "it's not required for someone to publish their list of trusted people", it's not even required for someone to publish what criteria they use
#
tantek whether it's a static list
#
tantek or a dynamic list like "people they follow on Twitter"
#
tantek and that is deliberately *by design*
#
tantek to make it harder for spammers and other attackers to attempt to "game" vouch
#
tantek the more diverse we have such vouch "approval" algorithms, the harder it is for spammers and anyone else to build anything for attacking that scales
#
rhiaro Sure, but do you just have to keep sending your webmention with random people as vouch and hope you get one?
#
tantek similar to species immunity through biodiversity
#
tantek rhiaro: no you have to start with the UI of the problem
#
tantek not the "what do I send?" (protocol) of the problem
#
tantek that's also by design
#
tantek it forces you, as the webmention sender, to think about the UI your software should display to you, the human, when a webmention is rejected, and a vouch is required
#
tantek another place that diversity helps
#
tantek these are all very deliberate aspects of the vouch design
#
tantek UI and human considerations first
#
rhiaro Okay
#
rhiaro Has anyone brainstormed this in detail yet?
#
rhiaro checks wiki
#
tantek brainstormed what in particular?
#
rhiaro the UI of what to do if a webmention with a vouch is rejected
KevinMarks_ joined the channel
#
tantek yes we did some brainstorming on that at IWC Cambridge 2014
#
tantek and ben_thatmustbeme built a UI for himself accordingly!
#
rhiaro aha, I see the wiki page is 'todo: update from IWC 2014 braindump'
#
rhiaro ben_thatmustbeme: is that visible somewhere?
#
tantek the respective IRC logs likely
#
tantek and maybe some bits on the video of the sessions
#
rhiaro righto
caseorganic joined the channel
#
tantek yes the discussions went faster than I or anyone else was able to write them up
#
ben_thatmustbeme and IWC Cambridge 2014 i built it to pull from my own DB and display a vouch URL if it found one, but I didn't handle more than that. I actually dumped the UI entirely in favor of just always sending a vouch if I can find one/have one stored, and if not, oh well
#
ben_thatmustbeme i also only accept webmentions that are either whitelisted or have a valid vouch, but all are stored so i can later review and whitelist anyone who didn't get through
#
rhiaro ben_thatmustbeme: if you send a webmention to someone who requires a vouch, how do you decide what URL to send as a vouch, that's likely to be accepted?
#
ben_thatmustbeme rhiaro, I store all incoming links to my site via headers info. i will check those sites for my link and see if it has rel=nofollow, if it does, i discount that as a possible vouch. at this point i have a list of pages that link to me without the rel=nofollow (they would be possible vouches)
#
ben_thatmustbeme when I want to send one, i start at the URL i'm sending to and check for all urls that don't have rel=nofollow. if any of those domains are in my DB, i've found my vouch url. If not, I try all rel=me links and repeat this process (only one level)
#
ben_thatmustbeme so i try their h-feed ideally
#
rhiaro okay, I see
#
ben_thatmustbeme if I don't find it after that, i just send a standard webmention and hope they are logging / will accept
#
ben_thatmustbeme eventually. I don't even know if they used vouch or not at that point
#
ben_thatmustbeme I also store that for every outgoing, so I don't do that search more than once per host
#
ben_thatmustbeme i still need to expire my possible vouches as they may have broken
#
ben_thatmustbeme but if they are accepting vouches, and they accept one from me, ideally i should be automatically approved from then on
#
ben_thatmustbeme this all explains why i created a link to ben.thatmustbe.me/whitelist in my source
#
ben_thatmustbeme so people could easily have a way to find my list
#
ben_thatmustbeme I want to move that whitelist to a friends list really
#
ben_thatmustbeme I don't actually display everyone in that list, i have public/private options in there
elijah and bblfish joined the channel