#social 2017-06-06
2017-06-06 UTC
timbl joined the channel
# 
ben_thatmustbeme ajordan++ thanks for the read through
# 
Loqi Sandro made 2 edits to [[Socialwg/AccountDiscovery]] https://www.w3.org/wiki/index.php?diff=103106&oldid=103078
tantek joined the channel
# Loqi Cwebber2 made 1 edit to [[Socialwg/2017-06-06]] https://www.w3.org/wiki/index.php?diff=103122&oldid=103075
tantek joined the channel
# Loqi Aaronpk made 1 edit to [[Socialwg/2017-06-06]] https://www.w3.org/wiki/index.php?diff=103123&oldid=103122
RRSAgent joined the channel
# 
RRSAgent logging to http://www.w3.org/2017/06/06-social-irc
Zakim joined the channel
# ben_thatmustbeme present+
# ben_thatmustbeme i can scribe
# ben_thatmustbeme scribenick: ben_thatmustbeme
# ben_thatmustbeme scribe: Ben Roberts
# ben_thatmustbeme Chair: tantek
# ben_thatmustbeme TOPIC: review of minutes of last week
# ben_thatmustbeme thanks ajordan
# ben_thatmustbeme RESOLVED: accept https://www.w3.org/wiki/Socialwg/2017-05-30-minutes of minutes from last week
# ben_thatmustbeme TOPIC: update for extension
# ben_thatmustbeme sandro: no update yet, i created a new w3c.social mastodon instance to hopefully show some support
# ben_thatmustbeme ... not sure how to show off that link
# ben_thatmustbeme tantek: well at least on the WG website
# ben_thatmustbeme sandro: i don't think that would be noticablel enough
# ben_thatmustbeme ajordan: twitter?
# ben_thatmustbeme sandro: oh yeah
# ben_thatmustbeme (laughs all around)
# ben_thatmustbeme TOPIC: specs
# ben_thatmustbeme tantek: with cwebber2 not here i don't know we can do much on activitypub
# ben_thatmustbeme ... anyone have any hot topics to raise on it
# ben_thatmustbeme ajordan: nothing has been raised recently in the issues
# ben_thatmustbeme tantek: we don't have aaronpk or julien here, so i don't know we can do much with websub
# 
aaronpk not on the call, but my only update is that pfefferle submitted an implementationreport for wordpress https://github.com/w3c/websub/blob/master/implementation-reports/PUBLISHER-wordpress-pubsubhubbub.md
# ben_thatmustbeme ... aaronpk did get some websub implementation report for wordpress
# ben_thatmustbeme ... its a plugin for worpress
# ben_thatmustbeme ajordan: i happen to have last weeks minutes open, and we were goign to discuss the direct message issue
# ben_thatmustbeme ... i just wanted to mention it because it seems like its been resolved
# ben_thatmustbeme .. it seems like everyone is on the same page
# ben_thatmustbeme tantek: sounds good, i leave it cwebber2 to prompt the original opener to ask if they are satisfied
# ben_thatmustbeme tantek: but good to see the updates there
# ben_thatmustbeme tantek: the only thing i have to report on post type discovery is in a little bit of messaging with sandro, i have had some ...
# ben_thatmustbeme got interrupted by co-worker and missed some there
# ben_thatmustbeme ... mastodon works well with other mastodon instances, but not much with others outside of mastodon
# ben_thatmustbeme sandro: some others have written plugins to make it work here and there
# ben_thatmustbeme tantek: does mastodon have a character limit?
# ben_thatmustbeme sandro: yes, but only on the input
# ben_thatmustbeme tantek: if we were to hlep make mastodon accept webmention, it would need to figure those out by post type discovery or a subset of the PTD algorithm
# ben_thatmustbeme ... i think i will look in to doing that subset, and it should be even easier to get more interop on it
# ben_thatmustbeme sandro: since we have such an open agenda, i wonder if we could talk more on that
# ben_thatmustbeme TOPIC: JF2
ajordan_ joined the channel
# ben_thatmustbeme oops
# ben_thatmustbeme i can help clean that up in the minutes ajordan
# ben_thatmustbeme sandro++
# ben_thatmustbeme q+ to talk about various parts
# ben_thatmustbeme not sure why my mic isn't working
# ben_thatmustbeme brb
# 
tantek aside: PTD issue filed for RTD https://github.com/tantek/post-type-discovery/issues/24
# ben_thatmustbeme ajordan++ for taking a hard day of scribing
# ben_thatmustbeme If a response to the source URL is shown on the source URL page (e.g. as a comment), then sender should treat that as an update of the source URL and re-send any previously sent Webmentions.
timbl joined the channel
# ben_thatmustbeme that was just a check box as i remember
# ben_thatmustbeme sandro++ good discussion
# ben_thatmustbeme ajordan++
# ben_thatmustbeme loqi doesn't like me anymore huh?
# RRSAgent I have made the request to generate http://www.w3.org/2017/06/06-social-minutes.html trackbot
# ben_thatmustbeme sandro: should I have announced https://indieweb.org/YAMLFeed as well? :P
# ben_thatmustbeme generally i do
# ben_thatmustbeme ajordan: hahahaha
# ben_thatmustbeme I still like the idea of EDIfeed
# ben_thatmustbeme no actual spec, just a corporate looking website with info on how to become a gold member
# ben_thatmustbeme and a contact form for pricing
# 
astronouth7303 hahaha +1 edifeed
# astronouth7303 YAMLFeed is clearly superior, and backwards compatible. All JSONFeed is automatically valid YAMLFeed
# ben_thatmustbeme astronouth7303: if I can make the spec work that way, i totally will
# ben_thatmustbeme definitely going to write it up eventually
# astronouth7303 ben_thatmustbeme: oh, is it more than just "take JSONFeed and run it through a YAML serializer"?
# ben_thatmustbeme Zakim, bye
# ben_thatmustbeme hey, there is implementation experience, i need to look at what fields those two existing implementation use
# ben_thatmustbeme no, not on that list
# ben_thatmustbeme mine isn't public yet
# ben_thatmustbeme saranix: he did mention on the call "oh i'll have to reply to saranix after this call"
# ben_thatmustbeme good catch ajordan
# saranix thx
# saranix ajordan++
# 
ajordan which, btw, are up: https://www.w3.org/wiki/Socialwg/2017-06-06-minutes
# ben_thatmustbeme can you make sure the #resolution link works as you edit it too
# ben_thatmustbeme oh, nevermind
# ben_thatmustbeme hadn't seen your line about them being up yet
# ben_thatmustbeme i got it
# ben_thatmustbeme the tool doesn't actually put IDs on the resolutions
# ben_thatmustbeme but for publishing a draft you need to link to the resolution, so its nicer to have https://www.w3.org/wiki/Socialwg/2017-06-06-minutes#resolution02 work correctly
# Loqi Strugee made 3 edits to [[Socialwg/2017-06-06-minutes]] https://www.w3.org/wiki/index.php?diff=103129&oldid=0
# Loqi Benthatmustbeme made 1 edit to [[Socialwg/2017-06-06-minutes]] https://www.w3.org/wiki/index.php?diff=103130&oldid=103129
# ben_thatmustbeme then every once in a long while i go an update this page
# ben_thatmustbeme for those that want to search through all past resolutions
# ben_thatmustbeme sometimes i leave out things like approval of minutes, as its not too important
# ben_thatmustbeme let me know when done, we missed a few meetings on that page
# ben_thatmustbeme and skipping some would be way worse than skipping them all
# ben_thatmustbeme as i wouldn't know which are missing
# ben_thatmustbeme awesome, thanks
# ben_thatmustbeme ajordan++
# ben_thatmustbeme karma fest
# ben_thatmustbeme i think i got them all
# ben_thatmustbeme i usually end up doing them in large batches
# ben_thatmustbeme huh??
# ben_thatmustbeme on what page?
# ben_thatmustbeme funky, if you edit and save the page (with exactly the same text) it fixes that
# ben_thatmustbeme btw, nice job keeping up with the logs, i thought it would be me mostly minuting. be glad you aren't doing a F2F
# ben_thatmustbeme if you minute those, you usually end up taking a chunk of several hours at a time
# ben_thatmustbeme but do know this, as scribe you are the only person who can stop the chair
# ben_thatmustbeme hehe
# ben_thatmustbeme yes, the F2Fs are fun indeed
# ben_thatmustbeme very hard when it gets to super technical areas
# Loqi Strugee made 1 edit to [[Socialwg/resolutions]] https://www.w3.org/wiki/index.php?diff=103131&oldid=103009
# ben_thatmustbeme cwebber2: you are pretty hard to scribe at times too
# ben_thatmustbeme haha
# ben_thatmustbeme the webex UI used to let you edit the names, so i would just edit names as people came on
# ben_thatmustbeme and so i could just look where the speaker icon was going nuts
# ben_thatmustbeme but it doesn't do that (on the web at least) any more
# ben_thatmustbeme at least i couldn't see it
# ben_thatmustbeme that is one of the advantages to mumble certainly
# ben_thatmustbeme s/to/of/
puckipedia joined the channel
# puckipedia hello
# ben_thatmustbeme ajordan: most of the people just call in, i'm usually one of only one or two that don't
# puckipedia well. my ActivityPub implementation seems to almost check all the boxes
# puckipedia yeah. Haven't implemented uploading media, and I haven't yet implemented any server->server authorization
# puckipedia it's not yet public, though. I do have it running publicly for some tests (I have built in OStatus federation)
# puckipedia I'm very happy ActivityPub requires IDs for basically everything
# puckipedia so far the implementation report looks like this https://gist.github.com/puckipedia/23b109c5ffafa522ec2bb49cee4c9c7b
# 
jaywink really hopes server to server authorization doesn't become a thing and instead authorship is validated in other ways (=signatures or remote lookup). Lots of good nice use cases ruined when servers don't just look at the payload but require some kind of auth. It doesn't even do verification of content authorship at all, so that needs to be done anyway.
# astronouth7303 i keep thinking that applying HTTP signatures to S2S comms makes a bunch of security much easier while still allowing web clients
# astronouth7303 especially since i think there's several kind of things that don't have remote lookups
# astronouth7303 mhm
# astronouth7303 oh, and i would feel much better about lookups if I knew who was asking
# astronouth7303 (limited audience, DMs, etc)
# astronouth7303 right
# astronouth7303 *most cases
# astronouth7303 like i said, i think there's some activities/objects that are ephemeral/don't get URLs. But I also need to read the specs again.
# astronouth7303 I advocate the use of HTTP signatures of all S2S, so you can verify who is making requests. You can apply it more tactfully for pushes (if it's the provider that owns the originating content or otherwise would be reasonable that they'd do the pushing)
# astronouth7303 but if you don't have the data at all, you're hands are tied, and you can't even make policies
# astronouth7303 and i consider it essential for non-push cases
# astronouth7303 yeah, if foo.com comments on something handled by bar.com, and spam.com is subscribed to notifications, you should expect to receive foo.com's stuff by way of bar.com
# astronouth7303 but if i'm spam.com and i get something from foo.com about bar.com from stranger.com, i should be viewing that with suspicion
# astronouth7303 *by way of stranger.com
# astronouth7303 yeah
# astronouth7303 it's probably a good policy that pushed content is verified when possible
# astronouth7303 i also think that LD signatures have a place, but I'm not sure what it is? Do you save it for client signatures? Can you have both server and client signatures? what about web clients?
# astronouth7303 when (if?) i implement my own server, i plan on just doing http signatures on everything
# astronouth7303 just make it part of the requests library
# astronouth7303 yeah, there's some really important issues that are left open
# astronouth7303 such as S2S auth and security
# astronouth7303 and federated identity
# astronouth7303 "locked down" is such a terse way of summarizing the huge tangled mess of making an open-access federated system secure
# astronouth7303 are there proposals besides HTTP signatures and LD signatures?
# astronouth7303 ewww
# astronouth7303 and also, we're not using any of the rest of JWT anywhere, are we?
# astronouth7303 add to the list of essentials: C2S auth
# astronouth7303 kind of an important piece if you want client software to work with multiple providers
# puckipedia I think that's semi-covered already
# astronouth7303 that is NOT how i would have ordered that?
# puckipedia the Actor object has a oauthAuthorizationEndpoint and oauthTokenEndpoint, which can be used to implement OAuth 2
# astronouth7303 last i read the AP spec, it was just suggestions, and the oauth endpoints were on the chopping block
# astronouth7303 also, if oauth is the "true path", there has to be some normalization about things like scope and a few other things
# astronouth7303 oauth2 is somewhat vague on a few details
# puckipedia tbh one of the At Risk features I'd preferably see completely scrapped is the object creation without a Create Activity.
# astronouth7303 oauth2 is good enough for C2S auth.
# astronouth7303 client oauth's with server, makes requests with token
# puckipedia cwebber2: well, I did kinda miss the fact that Announce was suddenly an allowed Activity too
# puckipedia I sent an Announce to my server to test... and it was wrapped in a Create
# astronouth7303 nothing is signed, but you have a reasonable certainty who made the request
# puckipedia well, I kinda don't do JSON-LD
# astronouth7303 puckipedia: you are in the wrong part of town
# astronouth7303 how long will that last?
# puckipedia I kinda think I do synchronize extensions properly?
# astronouth7303 i'm waiting for the first server that doesn't use the standard JSON-LD compaction algorithm
# puckipedia wait no, I can't really synchronize Objects if they are defined in non-default properties
# puckipedia well, there's kinda only one shitty JSON-LD implementation for C#...
# astronouth7303 i think LD is great for this use case
# puckipedia which is an auto-translated Java implementation actually
# astronouth7303 but i think any developer writing an LD consumer that doesn't actually interpret the LD markup is just running on prayers
# astronouth7303 of course, a lot of computer runs on prayers, so it might not be that bad
# puckipedia so kinda the API and everything is horrible. I kinda worked on a JSON-LD tool but I couldn't be bothered to finish it...
# astronouth7303 again, assuming all LD producers use the standard compaction algorithm
# astronouth7303 and implement it correctly
# astronouth7303 and the first developer that comes along and thinks they have a better compaction algorithm?
# puckipedia ... I swear the ActivityPub spec said that it had to be compacted
# astronouth7303 yes, part of the JSON-LD spec is standard expansion and compaction algorithms
# puckipedia tbh the biggest issue I have with JSON-LD is that it's a pain to implement in statically typed languages, imo
# puckipedia and tbh that even counts for the weird subset I currently do now
# astronouth7303 just as true of any other generic markup system
# puckipedia cwebber2: currently I have a simple class that is basically a wrapper for a dictionary of lists
# puckipedia astronouth7303: yeah. sadly.
# puckipedia also when reading any AP object into the database I 'flatten' it, and turn any subobject into its own object with ID etc (unless transient)
# puckipedia :D
# ben_thatmustbeme wilkie: what is your homepage URL, wilkie.io seems to be down
# puckipedia it shouldn't be too hard to add JSON-LD smartness into my flattening algorithm once I can be bothered to implement it
# astronouth7303 i would be looking at old-school explicit interface systems, where you have to request an interface from an object.
# astronouth7303 cwebber2: that's exactly how i would do it in python, although i have no idea how it would deal with name collisions
# astronouth7303 i suppose you could also do C-like function-subject stuff, too
# puckipedia let's build a tiny GUI to create a new Actor and then just show it to you all :)
# astronouth7303 cwebber2: in what?
# astronouth7303 ok
# astronouth7303 not really, no
# astronouth7303 i would imagine so
# astronouth7303 i'll have to look in to that, although i would appreciate breaking out the LD handling from the AS2 vocab?
# astronouth7303 since I think there's other related specs built on LD?
# puckipedia I know
# puckipedia but it was kinda absent from the Client->Server list of activities :P
# puckipedia I kinda forgot to add it to the List Of Activities
# puckipedia but it'd be nice if the server would just refuse to accept an Activity it doesn't know, instead of blindly wrapping it in a Create
# ajordan cwebber2: I wonder if we should make "subtype of Activity" a link in https://w3c.github.io/activitypub/#object-without-create
# puckipedia (also the behavior of Undo for Announce is not really defined, I mean, I could remove it from any inboxes, but hmm)
# puckipedia remembers when they accidentally put an `object` on a non-Activity
# puckipedia well, I'll probably get around to implementing proper JSON-LD
# puckipedia well, this might be more of an implementation detail, but I guess I remove the original Announce from the inbox so it isn't shown
# puckipedia that seems about right
# puckipedia actor | object | target | result | origin | instrument
# puckipedia well, an Undo for a Like I just keep the original object, and just add the undo later into the inbox
# puckipedia intransitive have no object
# puckipedia I think any AS vocab item is optional?
# puckipedia that *is* true
# puckipedia also I looked at Mastodon's AP implementation for a bit and... well
# puckipedia boosts etc point towards Activities, and the id for federated posts that are boosted point towards the internal representation, not the representation on the server it's from
# puckipedia a tiny bit
# puckipedia and only read-only
# puckipedia and, at least imo, not yet usable for federation with full AP servers
# puckipedia oh and the outbox won't load on people with more than a few thousand posts
# puckipedia https://github.com/tootsuite/mastodon/issues/3439 might be a sliiiight DoS issue
# puckipedia they are
# puckipedia but they load all the public posts into the redis cache, then only use a small subset
# puckipedia like, the first and last, to determine first and last page IDs
# puckipedia though, https://github.com/tootsuite/mastodon/issues/3438
# puckipedia (yes, OrderedCollectionPage with normal `items` property)
# puckipedia bye
# astronouth7303 does LD have subclasses? or just interfaces?
# puckipedia there's subclasses, I think?
# astronouth7303 i've spent some, but i mostly just understand it as "remember RDF? Yeah, it's in JSON now"
# puckipedia wellllllll
# puckipedia AS2 defines subclasses, but JSON-LD is just types and properties and there's no mapping to which property in which type
# astronouth7303 i feel like AS2 should have interfaces (types) and no inheritence, just extensions
# astronouth7303 (ie, all Foo must also be of type Activity)
# ajordan various snippets from https://duckduckgo.com/?q=json-ld+subclasses seem to imply that there are ;)
# astronouth7303 i'll have to actually read LD, then
# astronouth7303 in my copious free time
# astronouth7303 maybe i should become a monk and dedicate myself to FOSS. i would have much more time to work on this, then
# astronouth7303 problem is capitalism
# puckipedia heh, I've got holiday after high school finals
# astronouth7303 yeah.... i'm about a decade too late for my gap year
# puckipedia I am thinking of getting a holiday job though
# puckipedia okay, my ActivityPub implementation is ready for semi-testing https://lol.puckipedia.com/auth/login -
# puckipedia just type in any username/pass
# astronouth7303 ajordan: capitalism
# puckipedia then you can create an Actor, and get a OAuth2 token for it (this is because this is the least code for testing)
# ajordan (shameless plug) the Recurse Center is particularly great for this: https://recurse.com
# puckipedia I had to put it on a subdomain with https so I could test federation with Mastodon etc...
# astronouth7303 it's on my todo list to write a single-user server for myself. But i also need to read some game rules, and finish a rewrite i'm doing for another project
# puckipedia first thing in mind is 'lol'
# puckipedia astronouth7303: tbh I kinda built this server partially for myself
# astronouth7303 oh, and i like having this thing called a social life
# puckipedia that's why the users and actors are seperated. one login account can have more than one actor linked
# astronouth7303 puckipedia: my goal is to make "https://astro73.com/" my mastadon handle
# puckipedia I... don't think that's possible
# astronouth7303 users are referred to by URL, right?
# puckipedia I read through mastodon rules, and the handle is always @[preferredUsername]@[domain]
# aaronpk same, with https://aaronpk.com
# astronouth7303 are they using webfinger naming?
# puckipedia yeah for webfinger
# astronouth7303 GUYS WE NEED TO NORMALIZE: EMAIL OR URL STYLE USER HANDLES??????
# astronouth7303 BECAUSE THIS IS FREAKING REDICULOUS AND JUST LOOKS BAD
# puckipedia OStatus can do both (With one sidenote for Quitter and '/user/' in your path)
# astronouth7303 ...
# astronouth7303 this is basic stuff.
# puckipedia yes.
# puckipedia it was a PITA to find this even
# astronouth7303 and looks really really bad
# astronouth7303 i don't strongly care (although I think that URL pointing at the agent is the simpler approach) as long as we have consensus
# puckipedia astronouth7303: yeah. URL for agent seems to be good imo
# puckipedia especially because that's a proper AP ID too
# astronouth7303 aaronpk: probably
# astronouth7303 is this something that we can discuss at the CG tomorrow?
# puckipedia yeah. and having e.g. host/user point towards a human-readable HTML view
# astronouth7303 ajordan: if we're going email style, possibly
# astronouth7303 although i suspect i'll break some users and servers with "@astro73.com"
# puckipedia MENTION_RE = /(?:^|[^\/[:word:]])@([a-z0-9_]+(?:@[a-z0-9\.\-]+[a-z0-9]+)?)/i
# puckipedia basically, @([a-z0-9_]+(@[a-z0-9\.\-]+[a-z0-9]+)?)
# puckipedia aka you have to have at least one char for it to be a mention
# astronouth7303 and i haven't read the webfinger spec, so i don't know how much it aligns with the email address RFC
# astronouth7303 because hahahahahaha fully-conforming email address validation
# puckipedia I kinda keep the preferredUsername as a mapping to webfinger, so OStatus federation uses the preferredUsername
# puckipedia oh, and things I was wondering (maybe with an extension): Mastodon uses the atom entry summary as a content warning, which I currently translate into AS2 summary. Probably most native
# puckipedia ... woops, inboxes and outboxes weren't properly tagged, that's why they didn't work
# puckipedia right. got federating follows to work again (from mastodon to my own server) - it's time to go to bed here
# puckipedia server will shut down as it's on my desktop
# puckipedia this was very productive, thank you all :)
# puckipedia ajordan: btw I tested the follow with your(?) test account on my server - https://lol.puckipedia.com/users/inbox/strugee?from_id=994
# puckipedia (mastodon is local, won't I just redirected it to localhost)
# puckipedia ... the unfollow didn't federate lol
# puckipedia yeah. the unfollow salmon msesage in Mastodon uses the same ID as the follow
# puckipedia I'll stay here for a while longer though :P The spec still isn't done, and neither is my server! :)
# astronouth7303 So I guess I should make a list of webfinger vs ld vs other?
# puckipedia for user resolving/mentioning/etc?
# astronouth7303 For canonical unique user references
# astronouth7303 I'm assuming that in-text mentions will use some kind of a non unique or provider specific thing. Honestly, as long as the post is tagged with mentions in general, I think humans can sort out which bit is whom
# puckipedia yeah. Mastodon's mentions and hashtags are really wonky - they have a seperate hashtag list, and hashtags in the post are just <a href>
# puckipedia same with mentions
# astronouth7303 Seems reasonable
# astronouth7303 But if I want to give someone my fedisocial handle, what do I say?
# astronouth7303 Does it vary by what provider you're typing it into?
# astronouth7303 Right
# astronouth7303 Not really meaningful with mass federated systems
# puckipedia in the case of Mastodon, you have to search for the handle
# puckipedia Qvitter has a small search bar where you can enter handles too
# astronouth7303 I don't know what the blog-oriented specifications of this group are based on
# astronouth7303 I think that being able to say astro73.com/jamie is a pretty critical use case
# astronouth7303 Especially in the face of user privacy
# astronouth7303 Eg, not every instance will allow you to search for users, or will list everyone for strangers off the street
# puckipedia iirc Qvitter first tries to parse it as an Atom feed, then tries to find links to an atom feed, then tries to webfinger something
# nightpool astronouth7303: mastodon convention is to use @user@instance.tld
# nightpool Which I think is better then URLs for a few different reasons.
# nightpool Mostly having to do with how people parse URLs vs how they parse handles
# puckipedia sleep is for the weak. I fixed the unfollow support https://lol.puckipedia.com/users/inbox/strugee
# puckipedia another fun discussion: server-side ActivityPub relaying, or client-side? aka should the server get any object for the client, or should the client get them all itself? tbh I would prefer server-side relaying with an option for client-side, as I can't really relay non-URLs to e.g. Mastodon
# puckipedia anyways now that unfollows work it's really time to go to bed
timbl joined the channel
# astronouth7303 nightpool: I don't care about the debate nearly as much as I care about the consensus
# astronouth7303 If we webfinger always, then do that
# astronouth7303 If we look at it for ld, then do that
# astronouth7303 There's a ux argument to be made for both sides
# astronouth7303 some software has better UX than others
# astronouth7303 it's not a myth, but we don't have a lot of control over it
# astronouth7303 appears to be
# ajordan definitely: https://status.github.com/
# astronouth7303 ajordan: i mean, yes, we should accept lots of things.
# astronouth7303 but we can't be sure every provider will accept anything, and disagreement in the community will just create user confusion
# astronouth7303 the idea is that there will be dozens of packages powering hundreds of providers, right?
# astronouth7303 we can't be certain that some obscure one won't implement webfinger, or implement it in a broken way, or w/e
# Loqi Major service outage. https://status.github.com/
# astronouth7303 or some semi-corporate one
# astronouth7303 or w/e
# astronouth7303 or any other optional spec
# astronouth7303 on top of that, it just looks bad (and appearing like we have it together is important to draw the masses) if we can't agree on the form of user handle
# astronouth7303 wilkie: does this UI work across providers?
# astronouth7303 that's kind of the point of federation?
# astronouth7303 it doesn't matter if i'm not on the same provider as my friends
# astronouth7303 basically, yeah
# astronouth7303 but your local contact list is pretty limited, and everything assumes you'll be typing in things that aren't in it
# astronouth7303 that's not exactly the only way humans discover their online identities though
# astronouth7303 yeah, activitypub is all about LD and URLs pointing to LD
# astronouth7303 so what should we consider the canonical identifier?
# astronouth7303 that's what I would say
# astronouth7303 but that's not what's implemented, and there is dissent
# astronouth7303 if, say, mastodon were to display it's concept of the canonical, correct identifier for a user on their page, what would it be?
# astronouth7303 gnu social? diaspora?
# astronouth7303 friendica?
# astronouth7303 so a url
# astronouth7303 also, i thought activitypub suggested HTTP Content-Type negotiation?
# astronouth7303 ap doesn't mention it
# astronouth7303 there's been some revisions since I last looked at ActivityPub, but i don't recall any mentions of <link rel> as methods of agent discovery
# astronouth7303 and i'm fairly sure that it suggests content negotiation
# astronouth7303 i was mentally planning everything to be content-negotiated
# astronouth7303 "for federation" means "my user handed me this thing"
# astronouth7303 ie, it's not hidden from the user
# astronouth7303 ok, question 1: Are webfinger addresses normal or legacy?
# astronouth7303 2: Where is <link rel> discovery for ActivityPub specified?
# astronouth7303 so while conforming clients should accept webfinger addresses for compatibility, this is considered legacy and shouldn't be encouraged?
# astronouth7303 so half of FediVerse will display webfinger, and the other half will display URLs.
# astronouth7303 this is going to be great. A+
# astronouth7303 1. Stop trying to not own this. You know that 90% of clients will be the web UI from the provider
# astronouth7303 ajordan: it might not be part of ActivityPub, but there needs to be consistency if you want the ecosystem to take off
# astronouth7303 the mastodon i've seen is URLs, but that's like one guy
# astronouth7303 displaying/encouraging, or just accepting?
# astronouth7303 the point of the question is UI
# astronouth7303 what will common users recognize as their handle? What should they publish on business cards and websites? What will they tell their friends?
# astronouth7303 ok
# astronouth7303 do the mastodon implementers agree with you?
# astronouth7303 or any of the other implementations?
# astronouth7303 that's a necessary step to support webfinger at all? but is their attitude "here's the new workflow forever" or "yeah, this is mostly for transition reasons"?
# astronouth7303 that's my thing
# astronouth7303 are webfinger identifiers the way of the future or legacy comaptibility
# astronouth7303 and the other question, wilkie, where did <link rel> discovery come from?