#social 2019-04-18
2019-04-18 UTC
bblfish joined the channel
# 
dansup it's been so long since I touched the AP code in pixelfed, if https://www.w3.org/ns/activitystreams#Public is in cc its considered unlisted right?
# trwnh dansup: there's no official difference in delivery for to/cc, it's just about "primary" vs "secondary" audience. i personall interpret to: as "who is being targeted by this message" and cc: as "who has permission to see it". just like email. fwiw mastodon uses cc:public to indicate unlisted and to:public to indicate fully public.
# dansup perfect, thanks trwnh
# trwnh more context https://www.w3.org/TR/activitystreams-vocabulary/#audienceTargeting
bblfish, vitalyster, xmpp-social, puck and timbl joined the channel
# fr33domlover o/
# fr33domlover cjslep[m], have you by chance seen any usage of 'streams'? I'd like to add a new collection to my AP actors, I could use a top-level property, but was looking at 'streams' except I don't understand what its RDF datatype is
# fr33domlover AP spec says list, but, if it's just a list of URIs then it's not useful to me (I need it to be a dictionary, to have a distinct property for each collection)
timbl, _xmpp-social, benaiah, bblfish and Guest84 joined the channel
# trwnh fr33domlover: it doesn't seem like it has to be a list datatype -- 'streams' has a type of @id which means it can be any valid json-ld node i think. so you could use a nested json dict instead of an array. but you could probably also use 'endpoints' for anything "useful for this actor or someone referencing this actor"
# trwnh in practical terms i don't think 'streams' should be a pure list of uris without any other description, unless you intend implementations to dereference each of them and analyze the 'name' of each resulting Collection
# trwnh assuming the Collection was named, of course -- this is all left up to each implementation. but it doesn't seem useful to use un-named collections for 'streams' because the implication is that they should be "of interest" and therefore labeled somehow
bblfish and timbl joined the channel
# fr33domlover trwnh, I agree with your points, I'm just pointing out that the spec fails to explain the datatype of 'streams', and the context and old outdated OWL don't give info either, and the AS2 vocabulary spec doesn't list the streams property at all (it doesn't list endpoints either)
# fr33domlover So I'm basically complaining haha
# fr33domlover And I wish people by default defined vocabulary with RDF
# fr33domlover Because that makes sure you don't miss such details
# trwnh fr33domlover: 'streams' is part of the activitypub extension to as2, so of course as2 vocab doesn't list it
# fr33domlover trwnh, ah i see. well, i wish it was properly listed *somewhere* :p
# trwnh fr33domlover: https://w3.org/ns/activitystreams
# fr33domlover trwnh, yeah but there's no info there
# fr33domlover I don't mean literally list; it's listed in the AP spec too
# trwnh fr33domlover: it's linked to the spec it came from
# trwnh @id just means any json-ld node with an id
# fr33domlover I mean there's no explanation of the property's domain and range
# fr33domlover And no example
# trwnh yeah that's fair
# trwnh it's a very loosely defined extension
# fr33domlover trwnh, @id means any RDF resource, it basically tells us nothing
# fr33domlover It's just a JSON-LD feature to make it more compact
# trwnh what would the domain/range even be for 'streams' though
# fr33domlover "x":"https://y" instead of "x":
{"@id":"https://y"}# trwnh theoretically open to interpretation
# trwnh i do think that it would be really helpful if the activitypub spec documentation included the same level of definition that as2-vocab and as2-core do
# trwnh but as it is right now, 'streams' could be anything
# trwnh and no impl has used it yet afaik
# trwnh just some scattered discussion about using it for arbitrary subsets of 'outbox'
timbl, bblfish, dag[m] and etjet[m] joined the channel