#social 2019-03-26
2019-03-26 UTC
gazo2, jaft, ben_thatmust and vitalyster joined the channel
# 
fr33domlover Is the following behavior reasonable? Problem: When I comment on a ticket, I'd like to be able to address everyone who already commented on it. AP spec says the client does the addressing, so how do I determine my "to" and "cc" lists? Suggested solution: Everyone who comments on a ticket automatically becomes a ticket follower, and then I simply address the ticket's followers collection
# fr33domlover (I guess on the Fediverse, which currently doesn't rely on C2S, it works by doing addressing on the server side?)
# fr33domlover (Or the API can return a list of commenters to address?)
xmpp-social joined the channel
# cjslep[m] Just set 'to' to the tickets Collection of respondents. Only deliver to the ticket. Let Inbox forwarding handle the rest.
mrflos joined the channel
# cjslep[m] fr33domlover and in the C2S scenario the C side just puts the "respondents" target collection in it's "to" and the S side when it does it's separate bookkeeping of actual recipients will be able to expand that to actors because it owns that collection, so S can then deliver.
# fr33domlover cjslep yeah that's what I mean :) I'm asking if its okay/good to name that collection "followers", even though tickets aren't currently actors
# fr33domlover And likely won't be, I guess
# dansup hey fr33domlover
# dansup what do you think of the term: Pixelfed Lens? What does it remind you of?
# dansup As they say, naming things are hard
# dansup nvm
# dansup "Without a lens, everything is blurry." oh man, I can't wait :D
# fr33domlover Hey dansup! PixelFed Lens makes me think of some community focus feature, like some popular/pinned/featured posts. Or perhaps some image processing feature that allows to make pictures appear with visual effects
# fr33domlover dansup, am I close? ^_^
# dansup Pixelfed Lens hopes to replace your default camera app on Android, and integrate with Pixelfed Camera Roll for seamless continuity across devices. Take a Lens photo and edit it on desktop using Pixelfed Photographer Mode!
# dansup fr33domlover: pretty close ;)
# fr33domlover dansup someone close to me started using Instagram and I was thinking, hmm how could they be using PixelFed instead. But if the people you want to follow are there... what do you do :-/ I never used instagram myself , not even sure exactly what it is,but.. the lock in is sad :-(
# dansup not going to lie, coming up with brand names for every feature has helped users understand them a lot better. Keeping it simple is ideal.
# dansup yeah pixelfed isn't ready yet
# dansup IG does have a really handy data export tool that pixelfed uses for IG import
# fr33domlover dansup hmm what is PixelFed missing? Also I'm wondering, suppose I write a diary in handwriting and I want to publish it, by taking pictures of the pages, what are the differences feature-wise between using e.g. Mastodon for this, and using PixelFed?
# fr33domlover :)
# fr33domlover I guess I should try PixelFed :)
# fr33domlover Hmmm interesting, inbox forwarding says to look at inReplyTo, object, target and tag, but not context
# cjslep[m] Well, yes, at root someone is targeting the issue, no?
# cjslep[m] So recursively, up the tree you'll eventually find the issue. Satisfying that 1 criterion
# fr33domlover cjslep yeah sure. I mean, if someone targets the project, but not the ticket's "followers" collection, then the server will need to traverse inReplyTo recursively. But I also use context to specify the ticket being discussed, so instead of recursive inReplyTo expansion, I could grab the tree of replies without recursion
# cjslep[m] Sure, you can take as many shortcuts as you want. But I don't think it is a great idea. For instance, cross-mentioning tickets would break
# cjslep[m] Following the simple model of "a ticket is an Object", making that object a 'target' or 'tag' or 'inReplyTo', and implementing the inbox forwarding criteria should cover you. Doing anything else is putting in a lot of extra mental effort for not so great a payout, I think.\
# cjslep[m] Machine labor (executing a few more instructions) is cheap compared to human labor.
# cjslep[m] So while I agree 'context' is a great shortcut when it comes to easily fetching a conversation... I don't think it easily transitively applies to also being a shortcut to Inbox Forwarding, since "fetching a conversation" and "ghost reply problem" are two very different things to solve.
# cjslep[m] Plus... inbox forwarding has to generally be implemented for S2S anyway. It should be free to just plug this kind of Activity into that existing implementation.
# cjslep[m] Nothing special for tickets
timbl joined the channel
# fr33domlover cjslep is recursing through inReplyTo required? The spec says "and/or" as if you can choose which fields to examine. You can cross mention tickets using tag or whatever; the shortcut would only fail under "cross posting" i.e. more than one inReplyTo, but that scenario usually isn't supported for issue comments and perhaps I could reject it?
# fr33domlover But otherwise yeah, I could use plain inReplyTo recursion :)
timbl joined the channel
# xkr47 ooops old one
aaronpk joined the channel
# decentral1se go
# decentral1se woops hehe
GDorn, Simounet, aajualmok and timbl joined the channel
# fr33domlover cjslep, cwebber2, when I post a ticket comment I want to address 3 lists: project team; previous commenters; people subscribed to ticket updates. And also ticket author. Should I rely on client to address all 4 recipients (flexible but hmm less reliable?) or provide a pseudo collection that contains them all (safe and encapsulated, but kind of black box from client side, addressing becomes up to the
# fr33domlover server in practice)
timbl joined the channel
# cjslep[m] fr33domlover: You're not going to be able to enforce all C2S implementations to address a special collection, so once you have 1 special collection it won't matter if you have 3 more for a total of 4.
# cjslep[m] Also I think you need to be careful with your thought process. Inbox forwarding is only for the "other peer" side of S2S, and not the S side of C2S.