#social 2015-05-27

2015-05-27 UTC
#
jasnell I've made a number of significant updates to the current editor's draft reflecting recent conversations, decisions to simplify, and reconciling with the accepted user stories. I'm ready to cut an updated working draft but I need folks to review. For next Tuesday's WG call, I will add an item on the agenda to propose updating the current working draft based on the current Editor's Draft. Please take the time to review the updated Editor's draft
#
jasnell before then:
#
jasnell http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2.html
#
jasnell http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2-vocabulary.html
#
jasnell http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2-context.jsonld
#
jasnell http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2.owl (non-normative)
tilgovi, jasnell, bblfish, KevinMarks, pfefferle, stevenroose, pfefferle_, the_frey, the_frey_, edhelas, bblfish_, Loqi, Arnaud, melvster, raucao, aaronpk, kylewm, rhiaro, ben_thatmustbeme, oshepherd, dwhly, tessierashpool_, ElijahLynn, trackbot, mattl, bret, tsyesika, tommorris_, bigbluehat, JakeHart, slvrbckt, wilkie, elf-pavlik, sandro and AnnB joined the channel
#
AnnB topic: Social WG: https://www.w3.org/wiki/Socialwg - Next meeting agenda: https://www.w3.org/wiki/Socialwg/2015-06-02 - logs: http://socialwg.indiewebcamp.com/irc/socialwg/today
#
AnnB darn that didn't work
#
AnnB I need to be reminded of the "topic" syntax
hhalpin joined the channel
#
hhalpin Hey everyone, the Social IG meeting is beginning now on #socialig if folks want to join
#
hhalpin Meeting agenda here: https://www.w3.org/wiki/Socialig/2015-05-27
harry and jasnell joined the channel
#
ben_thatmustbeme AnnB: you can't set topic from a command, its some other controls in the client application. usually clicking the topic line, or some other button
jansauer and johnbreslin joined the channel
#
johnbreslin Recent slides by Peter Mika at WWW, may be of interest: http://www.slideshare.net/pmika1/social-networks-and-the-semantic-web-a-retrospective-of-the-past-10-years
#
hhalpin Note that I briefed him before that talk for about an hour :)
#
hhalpin Hopefully folks in the workshop were excited to see standards here.
sandro, johnbreslin, jasnell_, bblfish and cwebber2 joined the channel
#
cwebber2 o/
#
cwebber2 tsyesika: jasnell_: http://pamrel.lu/979be/ scheme constructors api in progress
#
tsyesika looks
#
tsyesika woot!
#
Loqi woot
#
cwebber2 pretty readable I think
#
tsyesika cwebber2: looking good
#
tsyesika cwebber2: oh i need to give you my new email address
#
cwebber2 tsyesika: :o
#
cwebber2 tsyesika: why not email me your new email address :)
#
tsyesika oh what a good idea
#
cwebber2 :)
#
tsyesika sent
#
cwebber2 cool
#
tsyesika it's alright if you use my old one it'll still get to me for a few years but i am starting the migration
#
cwebber2 gotcha :)
#
tsyesika :)
#
tsyesika but yeah cool about you starting that!
#
tsyesika in guile no less
#
cwebber2 tsyesika: the one thing I don't really have in this
#
cwebber2 is how to resolve non-default-vocabulary things, and the json-ld stuff
#
cwebber2 that'll be a whole load of tricky
#
cwebber2 since there's no json-ld library
#
cwebber2 for guile
#
tsyesika you might have to write one :)
#
cwebber2 I looked at the python one and it was 5k lines long
#
cwebber2 but
#
cwebber2 it might be easier in scheme
#
cwebber2 because of pattern matching
#
tsyesika nods
#
cwebber2 I really wish Python had pattern matching now that I'm used to it in scheme, though I'm not sure how it would be handled syntactically
#
cwebber2 (plus code using pattern matching tends to rely on tail call elimination to become efficient, so...)
#
cwebber2 could be done in Hy, but otherwise :)
#
cwebber2 oh btw tsyesika
#
cwebber2 thought this morning
#
cwebber2 about transient activities and etc
#
cwebber2 and "activities you should be able to eventually garbage collect"
#
cwebber2 I wonder if you could do something like "urn:uuid:f122aeb8-048a-11e5-be4e-0018f3be7a55"
#
cwebber2 most activities I don't even think really need public URIs anwyay
#
Loqi does a happy dance!
#
cwebber2 tsyesika: or, such is my thought since we determined that in the ActivityPump spec the main reason IDs are needed on activities is for data de-duplication
#
cwebber2 tsyesika: what do you think?
#
tsyesika so instead of using URIs use UUIDs?
#
cwebber2 tsyesika: yes, for activity ids
#
cwebber2 not object ids (yeah yeah activities are objects, but you know what I mean)
#
cwebber2 WDYT?
#
tsyesika hm
#
tsyesika i mean i like that all objects have dereferencable IDs that you can look dereference to look up or even use PUT or DELETE or whatever
#
tsyesika i admit you rarely delete or update an activity
#
tsyesika cwebber2: one thing is when a server gets an activity it's suppose to dereference it to verify that it's correct
#
cwebber2 tsyesika: is it? is that in the spec?
#
tsyesika "it should dereference the id both to ensure that it exists and is a valid object, and that it is not misrepresenting the object. (In this example, Mallory is likely spoofing an object allegedly posted by Alice.)"
#
cwebber2 > it should dereference the id both to ensure that it exists and is a valid object, and that it is not misrepresenting the object. (In this example, Mallory is likely spoofing an object allegedly posted by Alice.)
#
cwebber2 hm it seems so
#
cwebber2 yep
#
tsyesika i thought it was but sometime i imagine things
#
tsyesika xD
#
cwebber2 hm
#
cwebber2 tsyesika: wait but is it talking about the object or the activity
#
tsyesika both i think
#
tsyesika activities are objects
#
cwebber2 tsyesika: I wonder how much the "same-origin" thing prevents the thing I was discussing about off-site endpionts
#
cwebber2 endpoints
#
cwebber2 for static sites
#
cwebber2 it's not really clear how much that same-origin encapsulates
#
rhiaro waves
#
cwebber2 hi rhiaro !
#
rhiaro I made it to remote Croatian village :) I'll be back in 15 to catch up on all this exciting social chat I seem to have missed!
AnnB joined the channel
#
cwebber2 cool rhiaro
#
AnnB thanks, ben_thatmustbeme, for setting the topic .. and explaining I can't do it with a command
#
AnnB (I thought I had done so before)
#
AnnB looks like I had the wrong url for the log file ...
#
AnnB BTW, who has the group photo from the F2F in Paris?
#
aaronpk ooh me
#
AnnB would be neat if you could post that on the F2F wiki page
#
aaronpk did i not do that yet?
#
AnnB (pictures are good for my trip report :-)
#
AnnB I don't see it
#
cwebber2 tsyesika: https://github.com/w3c-social/activitypump/issues/34
#
cwebber2 rhiaro: ^^^^-- maybe of interest to you as well
#
harry in terms of json-ld cwebber2, it would be intersting to see how far you can get without writing a JSON-LD parser
#
harry we don't want people to write them obviously if they don't exist off the shelf
#
harry In terms of same origin, if you need to selectively run JS from other origins, there's always Content Security Policy (CSP) and even CORS specs, both have good support.
#
cwebber2 I think same-origin concerns aren't as much as running js from other origins
#
cwebber2 but more trusting who wrote this activity
#
cwebber2 and etc
#
cwebber2 harry: I'd like to also see how far I can go without json-ld
#
cwebber2 harry: obviously if I stick to AS2 core vocab, it's no problem
#
cwebber2 obviously I can't parse something where the type doesn't have a full URI without checking some context though
#
aaronpk AnnB: http://www.w3.org/wiki/File:2015-05-04-w3c-f2f.jpg
#
cwebber2 so there could be a faux "we know about this context meaning FooActivity means http://example.org/ns#FooActivity" type structure
#
cwebber2 but that would be more or less hardcoding the structure of contexts you know about without providing that processing
#
cwebber2 harry: are there other ways to go about it without pulling in json-ld for parsing? I can't think of any easily
#
cwebber2 harry: and thx for the feedback
#
ben_thatmustbeme AnnB, depending on client you might be able to do it with /topic
#
harry Off top of my head low budget stuff you can 1) ignore non-default vocabulary or 2) Store them in a lookup table
#
harry We had to do 2) with some big data processing as we kept breaking triple stores a while back
#
elf-pavlik cwebber2, JSON-LD parsing or JSON-LD processing? as in http://www.w3.org/TR/json-ld-api/
#
cwebber2 elf-pavlik: psh, you're acting like I use terminology right ever at all ;)
#
cwebber2 processing I think is right elf-pavlik
#
elf-pavlik which of the JSON-LD algorithms you see need to use? http://www.w3.org/TR/json-ld-api/#toc
#
cwebber2 elf-pavlik: I don't really know
#
cwebber2 I'd have to look
#
elf-pavlik in many cases you can treat JSON-LD as 'just JSON' and don't bother to use any of JSON-LD specific processing
#
Loqi Aaronpk made 1 edit to [[Socialwg/2015-05-04]] https://www.w3.org/wiki/index.php?diff=84269&oldid=84039
#
elf-pavlik also for properties not defined in AS2.0 vocab or aliased in AS2.0 JSON-LD context you should get quite far with http://www.w3.org/TR/activitystreams-core/#compact-iris
#
elf-pavlik and you shouldn't need to use full http://www.w3.org/TR/json-ld-api/#expansion
stevenroose, AnnB_ and tantek joined the channel
#
tantek jasnell_: just saw https://github.com/jasnell/w3c-socialwg-activitystreams/commit/3b419419026b240fbb2d1624984f058253aa98ab - you removed all the examples, even the correct ones - was that intentional?
#
tantek it also makes "fixing" them much more work
#
tantek trying to understand your goal here
#
tantek seems odd to include so many other syntaxes *besides* the most popular HTML extra-semantic markup on the web
#
tantek If you're not sure about specific examples - why not add a warning to them? E.g. "warning: example needs review"
#
tantek also makes it easier to see what's left to fix
#
tantek anyway - that's my request - revert that edit, and add warnings to the examples you're not sure about. If you're not sure about which ones to not be sure about - feel free to add the warning to all of the microformats examples that I haven't submitted a patch for yet.
#
jasnell_ the goal is to have examples that make sense, given that there is no normative mapping defined yet, having those examples in there is misleading.
#
tantek well you just made it *more work* to have microformats examples that make sense
#
tantek so that's why I'm questioning that edit
#
tantek and suggesting an alternate path forward
#
jasnell_ also, I'm going to be refactoring how the examples are inserted into the doc... essentially externalizing them and pulling them in programatically, which will make it significantly easier to add/edit/remove
#
tantek then why not wait to refactor all the examples?
#
tantek and frankly, why do you think the microdata examples (for example) are any less/more valid than the microformats examples?
#
tantek that edit has the appearance of singling out microformats
#
jasnell_ because there are normative mappings between json-ld, microdata and rdfa.
#
jasnell_ but I'm considering refactoring those significantly also to improve and simplify (specifically, I may end up removing the microdata and rdfa examples)
#
tantek there are? has anyone checked any of them? is anyone publishing any of them? or are all they all theoretical automatic transforms?
#
tantek which by the way are *more* misleading IMO (theoretical automatic transforms)
#
elf-pavlik jasnell_, how about leaving (and soon pulling in) only JSON-LD examples to activitystreams-core and publishing note activitystreams-examples with various other syntaxes shown together there?
#
elf-pavlik harry, ^
#
tantek elf-pavlik: that would be reasonable too
#
jasnell_ what would be useful is a separate doc that shows the microformats mapping + examples
#
jasnell_ the microdata and rdfa examples can be removed as well
#
elf-pavlik we can move all non JSON examples to separate document to avoid causing any confusions
#
jasnell_ that leaves the json-ld and turtle examples in the doc, which I think it useful to make it clearer how the json-ld processes down
#
elf-pavlik action-34
#
trackbot is looking up action-34.
#
trackbot action-34 -- Pavlik elf to add explaination to the spec about multiple serializations used in examples -- due 2015-02-10 -- OPEN
#
trackbot http://www.w3.org/Social/track/actions/34
#
elf-pavlik i have note saved on wiki here https://www.w3.org/wiki/Socialwg/Social_syntax#Serialization
#
elf-pavlik jasnell_, for people not familiar with turtle it might not make things easier to understand (myself i *like* turtle) so maybe better just keep JSON only in core spec and vocab and move all the other serializations to a separate WG Note ?
#
elf-pavlik also for Microdata we don't have tests, so far I only have automated JSON-LD - Turtle in core spec and started JSON-LD - RDFa in core
#
tantek elf-pavlik: what do your automated tests look like?
#
elf-pavlik once we separate examples i will finish automated tests for both Turtle and RDFa
#
tantek I'm wondering how much work it would be to add JSON-LD - microformats for AS2 at least.
#
tantek at least for a few examples
#
elf-pavlik tantek, https://github.com/jasnell/w3c-socialwg-activitystreams/blob/master/tests/examples.js#L7
#
tantek elf-pavlik++
#
Loqi elf-pavlik has 26 karma
#
elf-pavlik newer version https://github.com/elf-pavlik/w3c-socialwg-activitystreams/blob/test-rdfa/tests/helper.js#L60
#
elf-pavlik and RDFa https://github.com/elf-pavlik/w3c-socialwg-activitystreams/blob/test-rdfa/tests/helper.js#L80
#
tantek looks like everything assumes parsing into a graph
#
tantek that may be more work :/
#
elf-pavlik yes, I only compare 3 RDF serialization and use RDF Normalization to do it :)
#
elf-pavlik http://json-ld.github.io/normalization/tests/index.html
#
elf-pavlik tantek, why do you see parsing into a graph and issue?
#
tantek more work compared to parsing into a tree
#
tantek which e.g. microformats2 is defined to do: http://microformats.org/wiki/microformats2-parsing
#
elf-pavlik tantek, i'll try to get back on my very early experiment https://github.com/elf-pavlik/mf2jsonld
#
elf-pavlik on my local version i already started converting log of this channel to JSON-LD version
#
elf-pavlik tantek, i don't understand why do you treat "children" in special way instead of just having a regular "child" property?
#
elf-pavlik similar to http://www.w3.org/TR/activitystreams-vocabulary/#dfn-items
#
tantek elf-pavlik: because there may be an actual p-child property for example that has an embedded h-card
#
tantek might make sense to actually recast all the XFN values as p-* or u-* properties for this reason
#
elf-pavlik from vocab perspective p- , u- etc. doesn't matter? it only insturcts the parser where to look for value?
#
elf-pavlik assuming h: prefix, you would have h:child property, or h:first-name
#
elf-pavlik https://github.com/rhiaro/mf2rdf/issues/2
#
tantek elf-pavlik: and that's misleading - as the h-* prefix implies an object
#
elf-pavlik ok, mf2: prefix than :)
#
tantek folks in the microformats community seem to have settled on "mf" as a shorthand, e.g. in names of parsers
#
elf-pavlik mf2:child & mf2:first-name
#
tantek right, or mf2 for microformats2 in particular
#
tantek though given the back compat we've maintained, mf: is probably fine
#
elf-pavlik mf:child & mf:first-name
#
tantek elf-pavlik: you should know better than to ever suggest "first-name" as an example ;)
#
tantek that's like a person-vocab 101 error
#
tantek i18n err at least
#
elf-pavlik point taken ...
#
elf-pavlik afk ~5min
#
tantek seriously - that mistake is repeated SO often in programming book examples that it's really sad
#
tantek oh wait
#
tantek you know what else repeats that mistake?
#
tantek http://dublincore.org/documents/profile-guidelines/
#
tantek as documented here: http://tantek.com/2015/079/b1/dublin-core-application-profiles
#
tantek cc: rhiaro
#
tantek this is why I can't take Dublin Core seriously
#
tantek search for "foaf:firstName"
#
wilkie poco even got this mostly right
#
wilkie oh hey, poco's spec actually exists again
#
tantek wilkie: indeed! because poco (smartly) based their work on vCard (same methodology as hCard)
#
wilkie true
#
tantek so much easier to make fewer mistakes when you re-use others' open standard vocab, even if its not in the "data model" or "syntax" of your preference
#
rhiaro jasnell_: just fyi in this post I have a few microformats/AS2 mapping examples http://rhiaro.co.uk/2015/05/generating
#
rhiaro (that's not the focus of the post, but might be interesting)
#
elf-pavlik tantek, re: "not in the 'data model' or 'syntax' of your preference" for next week i aim to update my website with /candear and /map which will allow negotiating for JSON-LD, Turtle, RDFa using mentioned in AS2.0 spec geo: and ical: as well as h-feed with h-event and h-geo or even GeoJSON and iCalendar ...
#
elf-pavlik s/candear/calendar/
#
elf-pavlik https://github.com/w3c-social/social-vocab/tree/master/activity/Subscribe
#
tantek is still confused by GeoJSON which seems more like GeomJSON
#
elf-pavlik aaronpk, rhiaro what do you think about coorinating work around spacial+temporal use cases like my old capture https://www.w3.org/wiki/Socialwg/Buddyradar_Use_Case
#
tantek buddyradar sounds like surveillance
#
elf-pavlik tantek, sounds like you like ACL requirement :)
#
tantek elf-pavlik: the wording in particular sounds creepy
#
elf-pavlik tantek, http://lanyrd.com/profile/t/tracking/
#
Loqi Pelf made 1 edit to [[Socialwg/Buddyradar Use Case]] https://www.w3.org/wiki/index.php?diff=84273&oldid=77819
#
tantek lol elf-pavlik yeah that's not good microcopy!
#
tantek what's weird is that they are pulling that info from Twitter, which uses "following", and then renaming it to "tracking" for no reason!
#
elf-pavlik IndieWeb community could already use 'reader' with map/calendar based on collections like https://indiewebcamp.com/2015/Guest_List
#
tantek possibly yeah - there's definitely been some explorations of treating a simple list of h-cards as a reading list
#
elf-pavlik h-event + h-geo + h-card (list) .... when, where, who
#
elf-pavlik as well as ical: + geo: + vcard:
#
tantek elf-pavlik: h-event is based on ical - let me know if you find differences
#
tantek h-geo and h-card are based on vcard
#
tantek same same
#
elf-pavlik melvster, bblfish can we try making SoLiD app for that with proper ACL ? ^
#
elf-pavlik how about h-event and iCalendar / ical: http://www.w3.org/TR/rdfcal/
the_frey joined the channel
#
elf-pavlik action: pelf Document aligement between h-event / ical: & h-card / vcard: & h-geo / geo:
#
trackbot is creating a new ACTION.
#
trackbot Created ACTION-67 - Document aligement between h-event / ical: & h-card / vcard: & h-geo / geo: [on Pavlik elf - due 2015-06-03].
#
tantek elf-pavlik: what is this separate "geo" you speak of?
#
tantek just saw the :
#
tantek geo: URL scheme?
#
tantek elf-pavlik: I'm not sure it's good practice to use the group's Tracker instance as your own personal to-do list
#
tantek remember that having everyone have to revM��"all your own such personal actions tends to slow the group down
#
aaronpk i'm pretty sure it's not
#
tantek elf-pavlik: perhaps if you ask Harry nicely he can set you up with your own tracker, which you could trigger with your own IRC channel even ;)
#
elf-pavlik http://www.w3.org/TR/rdfcal/ | http://www.w3.org/TR/vcard-rdf/ | http://www.w3.org/2003/01/geo/
#
elf-pavlik tantek, we discussed last week possibilities of aligning AS2.0 and Microformats in terms of used vocabularies, AS2.0 already includes previxes vcard: geo: ical: and we could have clear match with h-card h-geo h-event ...
#
tantek oh boy - it's worse than I thought, there's RDF geo: as well as URL geo:
#
tantek RFC 5870
#
tantek that's not confusing at all :)
#
tantek http://tools.ietf.org/html/rfc5870
#
tantek elf-pavlik: indeed - and that proliferation of prefixes is why OGP gave up on them
#
tantek it's too damaging to developer usability to be worth it
#
elf-pavlik prefix != scheme
the_frey_ joined the channel
#
tantek elf-pavlik: you can say that but when they both use the same series of chars "geo:" then confusion is inevitable, repeatedly
#
elf-pavlik i can see that possible confusion... for the record - http://www.w3.org/TR/curie/
#
cwebber2 tantek: totally unrelated, but not mentioned on call:
#
cwebber2 about the dual licensing, more or less the micropub stuff is already dual licensed via this route because it was CC0'ed before being submitted right?
#
cwebber2 seems like a nice procedure hack :)
#
tantek elf-pavlik: yeah - many folks have given up on curies because of this
#
tantek they're one of the most fragile aspects of namespaces
#
tantek cwebber2: it's one of the reasons to normatively reference and external spec (that is CC0) rather than copy/paste it wholesale into a W3C document
#
tantek it keeps the original spec with its original liberal licensing
#
tantek and yes - it's a very good approach (not even a hack!) because it's also modular
#
tantek reference *an external spec
#
elf-pavlik in JSON-LD one would use { ..., "location": { "geo:lat": 45.5048473, "geo"long": -122.6549551 } } but maybe could also { ..., "location": "geo:45.5048473,-122.6549551" }
#
tantek what no @context? ;)
#
elf-pavlik first one looks more like http://microformats.org/wiki/h-geo
#
elf-pavlik :D
#
tantek "location" is a property of h-event, and has been re-used in h-entry as well
#
elf-pavlik tantek, "is a property of ..." i thought one can use each property in any type?
#
elf-pavlik if any explicit type at all ...
#
elf-pavlik use any property on any instance
#
elf-pavlik needs to jump into *Cook* activity (or *cook* verb)
#
elf-pavlik i need to check out http://microformats.org/wiki/h-recipe
#
elf-pavlik tantek, thanks for the nudge for not using this tracker as my personal tracker, i'll calibrate :)
#
tantek thanks elf-pavlik appreciated :)
the_frey, tilgovi, bblfish and edhelas joined the channel
#
AnnB thanks aaronpk, for posting photo from F2F
#
AnnB I just added names ... please correct or add if I forgot someone (especially people who were on phone that I might have forgotten)
#
AnnB https://www.w3.org/wiki/Socialwg/2015-05-04
#
cwebber2 hey hey hey!
#
cwebber2 guess what works :D
#
cwebber2 http://pamrel.lu/feabd/
#
cwebber2 jasnell_: rhiaro: tsyesika: ^^^^
#
cwebber2 aaronpk: also, this might make making a microformats->as2 converter easier: ^^^
#
tsyesika ^_^
#
cwebber2 ohhh I forgot something in there :P
#
cwebber2 the @type
#
Loqi Abasset made 1 edit to [[Socialwg/2015-05-04]] https://www.w3.org/wiki/index.php?diff=84275&oldid=84269
#
cwebber2 soooo
#
cwebber2 I guess I need to make a context thing to include the types without the long-URLs for each one ;p
#
elf-pavlik cwebber2, let me know if i can help you with any aspects of JSON-LD context
#
elf-pavlik $ curl -H "Accept: application/ld+json" http://rhiaro.co.uk/activities | head
#
elf-pavlik "@context": [
#
elf-pavlik "http://www.w3.org/ns/activitystreams",
#
elf-pavlik { "me": "http://vocab.amy.so/blog#" }
#
elf-pavlik ]
bblfish joined the channel
#
cwebber2 elf-pavlik: thanks :)
melvster, Arnaud, the_frey, jasnell and the_frey_ joined the channel