eprodrom, dmitriz, rhunter, vitalyster and xmpp-social joined the channel
#rialtate[m]heluecht : saw your post about private groups. http signed gets are an easy way to authenticate private fetches and many implementations support it.
#rialtate[m]Also the proxyUrl property as I understand it is for c2s so clients can have servers mask their activity and also fetch private resources when the server holds their private key
timbl joined the channel
#cjslep[m]Another data point: go-fed processes an activity's side effects synchronously but the uses a dependency-injected interface to do the delivery, up to the library user to do it synchronously, asynchronously, however the library user has implemented it.
#jdormit[m]cjslep: I like that approach. Does go-fed handle signing the requests or does the library user have to do that themselves?
#cjslep[m]go-fed/http-sigs is available for those that don't want to roll their own.
#jdormit[m]I see. I'm taking a more opinionated/all-in-one approach - ideally it should just be few lines of code to integrate ActivityPub into a project usingy library. I think I'll go for a sensible default with a pluggable delivery system for users with advanced needs
#cwebber2eprodrom: interesting stuff happening there
#cwebber2eprodrom: there's a new recommendation track which would be "Evergreen Standards"
#cwebber2eprodrom: which would be edited in place over time, aka a living document or living standard
#cwebber2eprodrom: unlike the previous structure where you work all your way up, do PR, CR, etc, it will be continuously evolving
#cwebber2eprodrom: there's anticipating there will be kind of a crossover
#cwebber2eprodrom: some recs being on old rec track that might want to switch over to evergreen standards
#cwebber2eprodrom: there might be others that might make more sense to continue on REC track
#cwebber2eprodrom: so they see the idea of standards switching between tracks
#cwebber2eprodrom: that may sound crazy, but in our own situation we did a lot of work to get to our own stable version, and there's a lot of stuff in implementations around questions, comments, small non-breaking changes, etc
#cwebber2eprodrom: maybe in two to three years as we're not making many more changes it's time for ActivityPub 1.1 and we freeze it, it goes back in the old REC format
#cwebber2eprodrom: I think that's very interesting for our document formats
#cwebber2eprodrom: we've been trying to do something where the version on w3c.org might be different than our GH thing
#cwebber2eprodrom: so I wanted to bring the idea for discussion
#cwebber2eprodrom: there's a temptation when you write standards docs is to think that's how you solve it, that you solve it by writing more standards, but I think there's only so much you can do to in terms of editing standards
#cwebber2eprodrom: but I think we've had a problem where people say the spec says do it this way, but this other way is easier to do so that's what we're doing
#cwebber2eprodrom: and I think we've been like *shrug*
#cwebber2eprodrom: so in that sense I think we can better serve those people who are using these specs for reference
#cwebber2eprodrom: I think that if someone says "I'd like to make this change" and we say "sorry we're closed", that's not very helpful... so it may be helpful to have a structure where things are open for conversation
#cwebber2cwebber2: is this at the point where it's still in discussion?
#cwebber2eprodrom: yes I think so, and I'm not 100% sure I understand how it works
#cwebber2eprodrom: so it's definitely something where we need to look at it, then follow up and say "hey, we think these documents would be good candidates"
#cwebber2eprodrom: one I def think would def be a good candidate is AS2 vocab
#cwebber2eprodrom: giving more definition, showing what they do, etc etc
#cwebber2eprodrom: those are all good things to do and could be really helpful
#cwebber2eprodrom: AS2 core doesn't seem as volatile
#cwebber2eprodrom: so maybe that one doesn't make as much sense
#cwebber2eprodrom: but they talk specifically about vocabularies and registries
#cwebber2eprodrom: I think ActivityPub may also be a good one
#cwebber2cwebber2: recommendation then became sending {Reject {Follow}} at any future time
#cwebber2eprodrom: I feel like maybe the relationship is between three states, you request and it's undetermined, you accept, and then if you undo it set it back to undetermined, then you reject
#cwebber2eprodrom: it seems like what you've suggested is the cleanest state machine
#cwebber2eprodrom: yes, it's a clean state machine
#cwebber2melody: I think it might not matter either way
#cwebber2melody: I think there's not a distinction, because from Bob's perspective of being removed or undo'ed it's the same situation of not getting updates
#cwebber2melody: if I later sent a reject as well, you may be like I already unsubscribed from bob
#cwebber2melody: the one step unsubscribe is the same either way
#cwebber2cwebber2: I think we should probably accept default behavior?
#cwebber2eprodrom: I think it makes sense to accept the way mastodon is doing it
#jaywink[m]> <@irc_cwebber2:cybre.space> If we did a W3C workshop on federation and ActivityPub, would you want to attend? Is it something you'd be willing to travel for?