#social 2018-08-18
2018-08-18 UTC
xmpp-social and timbl joined the channel
# 
dhasenan Anyone have an example of a Group representation in activitypub? Trying to figure out how the member list being represented in the wild (if it is).
# nightpool[m] peetube does it
# nightpool[m] but I don't have any channels I can link to offhand
# JasonRobinson[m] peetube :D
# saranix I'm doing a group thing but the member list is always private. If it were represented it would just be a regular Followers collection though. Anyone who is a 'follower' is a member of the group.
# JasonRobinson[m] btw, hubzilla has groups too, but I wonder if they've modeled them in AP?
# saranix JasonRobinson[m]: same as I just described AFAIK
# nightpool[m] that sounds really weird
# nightpool[m] so noone can ever follow your group blog?
# nightpool[m] that would be totally inappropriate for most social network groups
# nightpool[m] > <@nightpool:cybre.space> that would be totally inappropriate for most social network groups
# saranix well depends on if you mean groups as in forums or groups as in ACLs. completely different things
# nightpool[m] hmm. I guess I'll amend that to *many
# saranix I use capabilities not acls so I can't comment on the latter
# nightpool[m] sure but NEITHER of those are groups as an actor
# nightpool[m] which is what Group means in ActivityStreams
# saranix nightpool[m]: sure it is. in my impl. I'm fairly certain Hubzilla sticks to a few basic Types but it is nothing more than a string difference
# nightpool[m] ig
# nightpool[m] I mean, I'm not saying there's a wrong way to do it
# nightpool[m] but "anyone who follows the group is a member" is really limited for many use cases
# nightpool[m] like group blogs or group video channels
# saranix only if you conflate permissions and semantics. They should be seperate. If you mash them, you end up with crappy results. It's the fault of the mashing.
# nightpool[m] shT
# nightpool[m] woah weird typo
# nightpool[m] *what?
# nightpool[m] how is "different people may comprise this group then follow it" conflating permissions and semantics?
# nightpool[m] I'm literally speaking entirely on a semantics level
# saranix lol
# saranix I don't really have time to explain. maybe later.
# nightpool[m] "thousands of people may follow a group of only six users" is ..... not a complicated usecase and plenty of platforms already have it
# saranix yeah. you're getting confused on the word "members". That's on you.
# saranix I'll try to make it simple, in the real world, all the people in the congregations are members of the church, but only certain people are clergy
# nightpool[m] afafshsjsj
# nightpool[m] that's.... not what I'm saying at all
# saranix it is what I'm saying
# saranix and as usual, all you ever care about is your own point of view
# saranix you tire me
# nightpool[m] if you don't want to argue about it that's fine, but you have to understand that "membership in a group" and "interested in the actions of that group" can be two completely different things for some use cases
# saranix and you have to understand that if you want to use different phrasing to describe something, that doesn't mean it has to function the way you imagine that phrasing to mean
# saranix words simply aren't that precise
# saranix especially english words, in a chat context
# nightpool[m] you were proposing a way of representing group membership that was very limited to your particular usecase
# nightpool[m] I pointed that out
# saranix dhasenan asked about other impls, I can talk about how mine does it, and how hubzilla does it, and I can guarantee that both of ours can handle these "edge use cases" you talk about, even though we handle membership as following. permissions are handled separately, completely separately. I don't know why that is so hard for you to understand.
# saranix you "pointed out" that my impl can't do that use case, but you're wrong, it can. Just because you don't understand how, doesn't mean it can't. You aren't pointing anything out except your own rigid thinking on the topic.
# nightpool[m] I was never talking about permissions, solely about representation
# nightpool[m] you're the one who brought up the topic of permissions
# saranix *sigh* and there is your problem. You refuse to think about permissions, but that's precisely what makes it work
# saranix I'm not saying it's the best way to do it, or the only way to do it. It's just the way my impl (and hubzilla, coincidentally) do it
# saranix because that's what dhasenan asked
# saranix so like I said, my impl uses capabilities. So certain members of the group have capability to create new objects of certain types, certain members can moderate, certain members can use shared attachment storage, etc. Following doesn't mean you have any particular standing in the group, just that you get activity updates. It's a bare minimum of "participating" in the group.
# saranix note: it doesn't even determine that you get *all* activity updates, that too is determined by your capabilities ("permissions"). It means only that you *can* get activity updates, because there is a federation pathway
# saranix so, if it were of interest, to publish a "list" of who the group "members" were, given the semantics that you want those words to mean, then someone who has permission to, would simply create a collection of those members, as it's own little object. that object can be named, or uri namespaced, or shared ("announced") to insterested parties, whomever and however you determine that to be, however your fancy desires.
# saranix each group can even do it differently if they want. and why shouldn't they? All groups (in the loosest human meaning of the word) are so varied
# saranix technically, there wouldn't even be anything stopping you from then adding link rel's to some kind of foaf semantics or something
# saranix and dropping it in the group actor "profile"
# saranix I doubt there would ever be a considerable number of consumers that would look for it a certain way though
# saranix so if you're looking for "one true way", there never will be one, ever.
# nightpool[m] k
# Gargron is alsoKnownAs greenlit yet?