2017-11-14 UTC
rowan, timbl and cdchapman joined the channel
# 04:40 Loqi [nightpool] Me and @Gargron and a couple other developers had a chat on Saturday, and we decided that this is currently the highest priority enhancement for us as a project. We hope to have at least partial fixes for this emerge soon, things that address some of... xmpp-social, cdchapman, bengo and JanKusanagi joined the channel
jankusanagi_, timbl and rowan joined the channel
# 15:08 puckipedia I know! it's only getting better btw :P now that I know what to have and what not to have # 15:08 puckipedia note that I'm also really building my own templating system now # 15:08 puckipedia I only just added support for nested object-getting inside one template, because it was getting absurd # 15:11 puckipedia also help people that announce OStatus posts crash my server lol # 15:12 cwebber2 puckipedia: I am glad you are having fun with it and building cool things! # 15:12 puckipedia working on a "reply" page # 15:15 puckipedia 226 | 2017-11-14 16:08:41.060096 | 2017-11-15 18:40:37.532654 | 23 | DeliverToActivityPubTask | {"TargetInbox":"https://puckipedia.com/inbox","ObjectId":"https://octodon.social/users/cwebber/statuses/99003407965803437/activity"} # 15:15 puckipedia I blame you # 15:16 puckipedia huh. mentioning a non-AP user also explodes # 15:16 puckipedia oh finally, the mystical non-upgrade mastodon # 15:24 cwebber2 you can hover over the names on the left for the description # 15:24 cwebber2 I wonder if you actually do have both of those but they were just added later so # 15:25 puckipedia I check if the attributedTo/actor of the object is equal to the actor for Update (and only local objects), and I trim down incoming objects to only their Origin (when pushed, I might not do this if you manually GET it, this is a bug) # 15:26 cwebber2 puckipedia: you also check http signatures iirc right? # 15:27 puckipedia yes. one small exception, because I don't have json-ld signatures: if you POST to the sharedInbox, the post has an http signature, *but* doesn't equal the actor in the object, it will remove its knowledge of the incoming object and only use the ID # 15:28 puckipedia this is for Mastodon's federated replies # 15:28 puckipedia just mentioning it :P # 15:29 cwebber2 that's two more components with 2 implementations marked "yes" # 15:30 cwebber2 suddenly realizing that server:security-considerations:filter-incoming-content is misfiled # 15:30 cwebber2 because we don't require (or even desire) that you "change" an incoming object # 15:32 cwebber2 puckipedia: oh I'm guessing you're getting close to being able to fill out the client section too, huh? # 15:55 puckipedia aaand I have replies working :P rowan joined the channel
# 16:02 aaronpk did you put an RTL control character in your name? # 16:02 puckipedia that's not me, but someone else, but yes # 16:48 puckipedia it's a miracle # 16:48 puckipedia Delete's actually Delete an object cdchapman and ajordan joined the channel
# 16:58 ajordan do we have a telecon today? my primary IRC bouncer is down :/ bengo, timbl, JanKusanagi and ajordan joined the channel
eprodrom joined the channel
# 18:01 eprodrom seems like it # 18:02 cwebber2 it says I'm the first one! is it true or did I dial into the wrong thing :) # 18:03 eprodrom can we get a scribe? # 18:03 eprodrom trackbot, start meeting RRSAgent joined the channel
# 18:03 eprodrom I'm going to wait till 5 after the hour # 18:04 eprodrom scribe please? # 18:07 Zakim Present: tantek, cwebber, npdoty, annbass, hadleybeeman, snarfed, torgo, rhiaro, ajordan, eprodrom # 18:07 Zakim ... cwebber2, adam, dlongley, DenSchub, jungkees, jaywink, raucao, saranix, bwn, erincandescent, jet, rhiaro, ben_thatmustbeme, Gargron, sknebel, melody, mattl, bigbluehat, # 18:07 Zakim ... surinna, bitbear, howl, dwhly, tsyesika, sandro, nightpool, trackbot, puckipedia # 18:07 Zakim On IRC I see RRSAgent, eprodrom, ajordan, JanKusanagi, bengo, cdchapman, rowan, xmpp-social, csarven, dlehn, flackr, KjetilK, hadleybeeman, Zakim, Chocobozzz, Loqi, aaronpk, er1n, # 18:07 ajordan it took less time than I thought to finish the poem I have due in an hour :P # 18:07 rhiaro there are a bunch of presents that aren't right # 18:09 ajordan two minutes after dialing in and I've already nerdsniped cwebber2 with something else yikes # 18:10 rhiaro eprodrom: Switching the agenda around. Starting with TPAC, then WebSub then AP tantek joined the channel
# 18:10 rhiaro ... can Chris and/or sandro tell us what happened? # 18:10 rhiaro sandro: I missed my flight, I wasn't at that part # 18:10 rhiaro cwebber2: It was more of a social CG gathering, but still interesting and relevant # 18:11 rhiaro ... Off the top of my memory, we started off with introducing people and what's happening in this space # 18:11 rhiaro ... including one of the mastodon node hosters was there, toot.cafe # 18:11 rhiaro ... we demo'd two mastodon sites interoperating # 18:11 Zakim Present: tantek, cwebber, npdoty, annbass, hadleybeeman, snarfed, torgo, rhiaro, ajordan, eprodrom, aaronpk, bengo, sandro # 18:11 Zakim ... aaronpk, er1n, cwebber2, adam, dlongley, DenSchub, jungkees, jaywink, raucao, saranix, bwn, erincandescent, jet, rhiaro, ben_thatmustbeme, Gargron, sknebel, melody, mattl, # 18:11 Zakim sees on irc: tantek, RRSAgent, eprodrom, ajordan, JanKusanagi, bengo, cdchapman, rowan, xmpp-social, csarven, dlehn, flackr, KjetilK, hadleybeeman, Zakim, Chocobozzz, Loqi, # 18:11 Zakim ... bigbluehat, surinna, bitbear, howl, dwhly, tsyesika, sandro, nightpool, trackbot, puckipedia # 18:11 rhiaro ... we then moved to discussing anti abuse tooling # 18:11 rhiaro ... and what we could do to get the ideas effective. a lot about distributed block lists and other things # 18:12 rhiaro ... and then it went into something that's not currently work in the socialcg but a favourite rathole of mine # 18:12 rhiaro ... what would it bel ike ifyou had a federated MUD or other federated game timbl joined the channel
# 18:12 rhiaro ... But people were excited about social spaces with rules # 18:12 tantek (we discussed Vouch a bit too as part of anti-abuse) # 18:12 rhiaro eprodrom: was there any discussion by any Members who showed up about either of the specs that we're working on right now? jankusanagi_ joined the channel
# 18:13 rhiaro cwebber2: there were some people from the TAG who were interested in this stuff # 18:13 rhiaro ... There were two separate meetings! I'm mixing them up # 18:13 rhiaro ... We had a conversation where the TAG showed up and I'm tryign to remember everything that happened there. We spent more time talking about .. we also had somebody from bridgy (snarfed Ryan Barrett) # 18:13 rhiaro ... that's where most of the conversation about the anti abuse tooling happened # 18:13 tantek eprodrom, there were questions about like should brands/companies just install and launch their own Mastodon instances? and we discussed the Monoculture problem # 18:14 rhiaro ... The TAG didn't know mastodon were using our protocol # 18:14 tantek (like if everyone "just runs Mastodon", then AP becomes the Mastodon API instead of an open standard) # 18:14 rhiaro ... The majority of the conversation was anti abuse tooling # 18:14 rhiaro ... it doesn't seem like the network is active enough that abuse is a problem # 18:14 rhiaro cwebber2: there's a LOT of active moderation on mastodon # 18:15 rhiaro ... a lot of mastodon users are pretty happy is because mastodon builds in a lot of tooling # 18:15 rhiaro ... there has been a culture clash between people from different communities # 18:15 tantek unfortunately had a conflict this morning, meeting at work # 18:15 rhiaro ... people are concerned about making sure it keeps working # 18:16 tantek small (very small) little launch today at work # 18:16 rhiaro eprodrom: Talking about tooling, you mean like spam filters? Or is it more like stuff that's built into particular networks like silencing and so on? # 18:16 rhiaro cwebber2: people are interested in all of that across the board # 18:16 rhiaro ... there's the general feeling that we have to explore a number of different approaches and it's not clear which arethe most accessible or that we might need a combination ot really get the job done # 18:16 rhiaro ... a lot of people are really interested in federated block lists and a web of trust and distrust # 18:16 rhiaro ... There was one more topic that was pretty big # 18:17 Loqi [nightpool] Me and @Gargron and a couple other developers had a chat on Saturday, and we decided that this is currently the highest priority enhancement for us as a project. We hope to have at least partial fixes for this emerge soon, things that address some of... # 18:17 rhiaro ... and is increasing because it looks like this is going to be the next major topic in mastodon # 18:17 rhiaro ... still a big issue that people are really intersted in # 18:17 rhiaro ... how to handle if your server goes down or you want to move for whatever reason # 18:17 rhiaro ... maybe you want to move becus eyou're not happy with the moderation policies, or you're not happy because.. or you want to host your own # 18:17 rhiaro ... These remain the two biggest conversations in the CG # 18:18 rhiaro ... interesting also to see that the people who showed up without prior experience also wanted to talk about these issues # 18:18 rhiaro ... a lot of stuff that does not necessarily seem to be standards questions # 18:19 eprodrom scribenick: bengo # 18:20 bengo eprodrom is there a cg meeting tomorrow? # 18:20 bengo cwebber2 no, next week # 18:20 bengo eprodrom there will be some things that are patterns, some are one-off APIs, other things that might be standardized # 18:20 bengo sandro my perception of mastodon is that the community there polices in a way that there are good people and bad people # 18:21 bengo rhiaro (take scribe back?) # 18:21 bengo sandro I happen to agree with their notion of good peopel and bad people, but not their policing # 18:21 bengo sandro obviously you could have a fork where someone uses a fork and deploy a node cut off from mian network # 18:21 rhiaro eprodrom: it's interesting to see how these patterns work out # 18:21 rhiaro ... I think those are things we can talk about at the CG # 18:22 ajordan eprodrom: wondering if we can do AS2 since I have to leave # 18:23 Loqi [tantek] hey hey we're in the SocialCG f2f at TPAC!!! # 18:23 rhiaro sandro: I'm a little confused about process here. There are two comments and I believe they are technically member confidential # 18:24 rhiaro ... I don't know how the group is supposed to respond to that # 18:24 rhiaro ... they both request the change but don't require it # 18:24 tantek sandro, I believe there is only substantive comment (HTTPS must vs should) # 18:25 tantek the other is editorial about the WBS system vs what the spec says # 18:25 rhiaro sandro: one is that HTTPS should be a MUST rather than a SHOULD # 18:25 rhiaro ... our reply to that is that we can't for back compat # 18:26 tantek we can even resolve on that if someone can create the issue in GH fast enough :) # 18:27 rhiaro aaronpk: I agree we don't have to require it because it uses the secret to verify the payload # 18:27 rhiaro ... HTTPS just provides privacy on the transport, which is not a guarantee that we are claiming to make # 18:27 rhiaro ... The spec should be concerned about successfully delivering notifications, which can be verified by HTTPS or the secret # 18:27 rhiaro sandro: but if you don't use https when you send the secret? # 18:27 rhiaro aaronpk: the hashing mechanism verifies the payload even over http # 18:28 rhiaro cwebber2: it's a SHOULD in AP because it might have some sort of use inside a corporate network with http # 18:29 rhiaro ... 2 is that you may use it with a scheme that's not http at all. At Rebooting WoT something about ids # 18:29 rhiaro ... URIs have schemes beyond http and https. I feel like that's a reasonable justification here # 18:29 rhiaro eprodrom: in AP they act as identifiers so that's fine # 18:29 rhiaro ... in websub we're talking about actual endpoints that are being used for posting and retreiving things # 18:30 rhiaro ... it sounds like the question is is the verifiability of http with extra parameters in websub sufficient that we don't need to make this a MUST # 18:30 rhiaro eprodrom: Or we can have a resolution right now # 18:30 rhiaro aaronpk: let's resolve it on the call then we can reference a link # 18:30 eprodrom PROPOSED: retain HTTPS as a "should" or "recommended" requirement in WebSub # 18:31 rhiaro sandro: I don't agree with the security claim but I do agree with the back compat # 18:31 rhiaro aaronpk: you don't agree that the hash is enough to verify the paylaod? # 18:31 rhiaro sandro: on the delivery it is but on the earlier part it isn't # 18:31 rhiaro ... it would be higher priority to make the hub do it # 18:32 rhiaro eprodrom: basically we say it is strongly recommended # 18:32 eprodrom "It is strongly recommended to use HTTPS for all requests." # 18:32 rhiaro sandro: the notification endpoint, there's not that much reason but still probably a good idea # 18:33 eprodrom RESOLVED: retain HTTPS as a "should" or "recommended" requirement in WebSub # 18:33 ajordan s/LOL/lol/ oops capslock makin me seem overenthusiastic # 18:34 rhiaro sandro: second issue is editorial regarding editors list # 18:36 eprodrom PROPOSED: add invited expert status for author and editors of WebSub ajordan_ joined the channel
# 18:37 eprodrom RESOLVED: add invited expert status for author and editors of WebSub # 18:37 rhiaro sandro: at this point it's up to director not the WG # 18:38 tantek uh, if they're employees of W3C members I'm not sure we can do that # 18:38 rhiaro aaronpk: still a couple of open issues which might result in editorial changes # 18:38 rhiaro sandro: we should resolve that we'd like these changes # 18:38 tantek sandro, perhaps ask if they could join the WG? # 18:39 rhiaro aaronpk: one review we got I split up into issues and we haven't talked about them yet # 18:39 rhiaro eprodrom: what are our options for incorporating those? # 18:39 rhiaro ... something we need to decide here or should we discuss as we go to PR # 18:39 rhiaro sandro: we should solve before we ask for transition to REC # 18:39 tantek we have to come up with responses to all the comments on the PR # 18:39 tantek including the issues that aaronpk split-up from that one mega-issue # 18:40 rhiaro aaronpk: it's a bunch of really small things, should be editorial # 18:40 sandro tantek, feel free, but I'd be surprised if they'd do it, and I don't see why it would matter at this point # 18:40 rhiaro eprodrom: there are 17 items. I'm not sure we have time # 18:40 tantek we need 1) answers to those issues (either by aaronpk, or discussed here), and 2) resolutions from the WG on those answers # 18:43 rhiaro i thought my last messages hadn't gone but they did # 18:44 rhiaro cwebber2: there are two open issues, one is a personal TODO so I'll move it out # 18:44 rhiaro ... The implementation reports for mastodon is also not an issue, we're just trying to wrap up their implementation reports # 18:44 rhiaro ... so no open issues left, just two things for the editors' sake that won't affect the spec # 18:44 rhiaro ... Good news, we have implementation reports page and 5 submissions, and 1 more pending # 18:45 rhiaro ... things that don't yet have two submissions ^ # 18:45 rhiaro ... the bad news is it's today and we still have these # 18:45 rhiaro ... the good news is I think if I took a week we could get them # 18:46 bengo I can probably add some stuff to distbin where it helps get to 2 # 18:46 rhiaro <rhiaro> I need to catch up my implementation too :s # 18:46 rhiaro cwebber2: I assume we want to wait til next week to move to PR # 18:46 rhiaro sandro: we can't move forward until we have time # 18:46 rhiaro ... bot for WebSub and AP we need a call next week # 18:47 rhiaro eprodrom: working backwards, if we approve PR next week will that count? # 18:47 rhiaro I thinkw e're fine cos sandro and I sneakily built in a little buffer? # 18:47 rhiaro sandro: technically we only have to go to PR by the end of the WG. It looks bad, but we can technically do it # 18:48 rhiaro ... to actually go to REC by the end of the WG maybe next week we can still do it # 18:48 rhiaro ... let's try for next week, but it's not the end of the world if we can't # 18:48 rhiaro cwebber2: I'm in favour of everything ready for PR next week # 18:49 bengo evan can make activitypubcoin # 18:49 ajordan_ if I go *really* hard I might be able to land AP in pump.io master # 18:49 ajordan_ don't want to commit though because I have some papers due lol # 18:49 rhiaro eprodrom: it doesn' thave to go to master AJ, just has to exist # 18:49 bengo I bought activitypub.com # 18:49 bengo where should I redirect it? # 18:50 Zakim sees cwebber, ajordan_ on the speaker queue # 18:50 bengo activitypub.rocks? # 18:50 sandro ajordan_, it sounds like you don't need to do that for next week, but instead plan to have that for the REC announcement # 18:50 rhiaro eprodrom: hold onto it and sell it when AP is huge # 18:50 bengo activitypub.rocks # 18:50 bengo will do this week # 18:50 rhiaro eprodrom: it looks like we have 10 minutes so ajordan_ are you here? # 18:51 tantek q+ to note concern about outstanding issues on WebSub # 18:51 rhiaro ... I'm almost done weith a webmention implementation and websub shouldn't be too far off # 18:51 rhiaro ... something automatically do the right thing. # 18:51 rhiaro ... AS2 is really close to shipping in pump.io # 18:52 Zakim tantek, you wanted to note concern about outstanding issues on WebSub # 18:52 rhiaro tantek: we have quite a few number as the result of comments on websub # 18:53 rhiaro ... I think we need to promptly have answeres to those that the group resolves on, either resulting in no changes or editorial changes, assuming there isn't an actual implementation problem # 18:53 rhiaro eprodrom: we agreed earlier that aaron and julien are going to look through these, identify ones that don't require WG input and then we will next tuesday have another meeting where we'll vote on the rest of them # 18:53 rhiaro sandro: can we plan for next week being 90 mins? # 18:54 ajordan_ I can't do longer than 60 minutes as usual but by all means go for it # 18:54 rhiaro tantek: Second thing.. because we're trying to wrap up our normative documents and there's a lot of activity going on, and stuff beyond in the CG # 18:54 rhiaro ... there are a lot of implementations that interoperate in ways that we ight not have fully covered # 18:55 rhiaro ... I think there's opportunity here if interest to close out the year with a few informative notes that are kind of capturing the state of interop of extensions or additons to existing specs # 18:55 rhiaro ... just want to put it out as something to think about # 18:55 rhiaro ... Thinking about treating PTD that way because it's evolving # 18:56 cwebber2 eprodrom: one thing the CG is discussing is what's missing, maybe we can talk about what can be done there and move to a note, or move that work to the CG # 18:56 cwebber2 tantek: I'd love to have the CG have a note about what topics are worth exploring # 18:57 cwebber2 eprodrom: ok great, I want to move to ajordan_'s topic of discussion... if you want to bring it up now before 2:00? # 18:57 cwebber2 ajordan_: AS2 doc says how to transform AS1 -> AS2 but says nothing about the vocabulary # 18:57 tantek in particular I'd like to see us publish a number of "interop NOTEs" where we document uses of our specs with 3+ more interop impls # 18:57 cwebber2 ajordan_: pump.io has to transform as1 to as2 as last step of response handler, but some of them have no equivalent in as2 # 18:58 cwebber2 eprodrom: my first response is if we need to do that we'll do an extension vocabulary... I'm pretty suprised we have stuff in pump.io that didn't make it into as2 but sometimes that happens # 18:58 cwebber2 ajordan_: I have no data on what verbs people are actually using, so I don't know if there's data out there for it # 18:59 cwebber2 ajordan_: my assumpiton is a lot of verbs were dropped due to low evidence of implementation # 18:59 cwebber2 eprodrom: why don't you and I talk about this, if parts of pump.io don't have a match we'll do an extension vocab # 18:59 cwebber2 sandro: I encourage you put them in the as2 namespace, that's what the CG is meant to do # 19:00 cwebber2 eprodrom: you had asked me about this earlier ajordan_, we'll deal with it soon # 19:00 cwebber2 eprodrom: no CG meeting tomorrow so we'll deal with this 1 on 1 ajordan_ # 19:01 cwebber2 eprodrom: we have a meeting same time same place but longer next week # 19:01 ajordan_ thanks all! hope the noisy hallway wasn't too bad lol # 19:03 eprodrom OK, thanks everyone # 19:03 Zakim As of this point the attendees have been tantek, cwebber, npdoty, annbass, hadleybeeman, snarfed, torgo, rhiaro, ajordan, eprodrom, aaronpk, bengo, sandro # 19:05 eprodrom places.pub is a server that exposes places in the physical world as AS2 objects at an HTTPS URL, as AP requires # 19:06 eprodrom Yeah, it's getting there # 19:06 eprodrom There's also a search interface # 19:06 eprodrom So you can look up places in a bounding box and with a particular name pattern # 19:08 eprodrom I had to set up my own OpenStreetMap API server because apparently things like "searching by name" are too gruesome for the public API servers # 19:09 eprodrom And the search is sloooooowwwwww # 19:10 eprodrom With places.pub and tags.pub we start having some AS2 services that can be used for naming things globally # 19:10 tantek hmm, maybe too late? will leave it to cwebber to fix it in post cdchapman joined the channel
# 19:15 bengo does/will places.pub receive notifications in an AP inbox? # 19:16 puckipedia bengo: just for reference, my server really doesn't like not having an ID on the object, and/or IDs not being http # 19:16 eprodrom bengo: the goal is yes # 19:16 bengo "distbin:activityPubDeliveryFailures": [ # 19:16 bengo "name": "TargetRequestFailed", # 19:16 eprodrom bengo: same with tags.pub # 19:16 eprodrom bengo: what Accept header are you using? # 19:17 puckipedia my guess is application/ld+json with the # after the activitystreams # 19:17 bengo yeah I probably do, I implemented from like spec form a year ago # 19:17 eprodrom oh, with a profile defined? Hmm # 19:18 eprodrom so places.pub checks for activity+json, ld+json or just json, and after that it gives up # 19:18 eprodrom But if it's ld+json with a profile, I might not be checking that well # 19:18 puckipedia bengo: yup. I'm going to fix the inbox thing so Kroeg'll deliver to anything with an inbox # 19:19 puckipedia but stuff like replies being an inline collection, object not having an ID, and the activity not being a proper ID are things that my server won't support # 19:20 puckipedia (also you miiight want to define a proper context, e.g. ["https://www.w3.org/ns/activitystreams", {"distbin": "https://distbin.com/ns#"}] # 19:20 bengo puckipedia how is that improper? # 19:20 eprodrom Yeah, that should be OK # 19:21 eprodrom string, array, or object is OK for @context # 19:21 puckipedia well, I mean, it's a proper ID, but it isn't dereferencable # 19:22 bengo I'll make it a DID to make tantek so happy # 19:22 tantek apparently they're all the rage now, or you would think from the blockchainers at TPAC rowan joined the channel
# 19:22 puckipedia bengo: imagine if I said "to": ["urn:uuid:bb259a57-0c04-4f3b-acc0-589204cd8209"] # 19:22 puckipedia how would you deliver to that # 19:23 tantek I just figure it's the hype TLA of the year, like DHTs were last year # 19:23 bengo puckipedia you can look at .url # 19:23 bengo puckipedia I'll switch to https id's sometime # 19:24 puckipedia bengo: and please add a separate ID to the object and replies # 19:26 puckipedia ... my single-user AP instance already knows 353 users # 19:26 puckipedia nope there's duplicates timbl joined the channel
timbl and bengo joined the channel
cwebber2 nightpool: Gargron: cool to read https://github.com/tootsuite/mastodon/issues/177#issuecomment-344056590 ... I'm really interested in helping to see if we can find a good direction, assuming you'd welcome my input on the direction you'd go.
puckipedia https://dl.puckipedia.com/up/ooh_fancy.png this looks so niiiice
Loqi [Flammekueche Sournoise ✨?☕] oh. my. god. "behind the scenes it is using screen" omg please help https://github.com/sebbekarlsson/systemctl
RRSAgent logging to https://www.w3.org/2017/11/14-social-irc
tantek SocialCG meeting at TPAC: https://www.w3.org/wiki/SocialCG/2017-11-06