#social 2019-04-07
2019-04-07 UTC
naturzukunft, xmpp-social and vitalyster joined the channel; vitalyster left the channel
# 
jdormit[m] Is it valid for a client to address an activity to a Link object and expect the server to resolve the Link into an actor and deliver it?
# rialtate[m] Like rel=me? Don't think anything supports it and sounds very strange to me. How did this come up?
biodan joined the channel
# jdormit[m] Like the ActivityStreams Link object type
# jdormit[m] https://www.w3.org/ns/activitystreams#Link
# biodan For this summer I intend to implement pixelfed-like stories on GS, any suggestions on how it should work? (cc: dansup)
# rialtate[m] jdormit: yes but why would you address something to a Link object? What does that even mean in human terms?
# nightpool[m] a Link is not an Actor
# jdormit[m] Hmm yeah I think I'm overthinking stuff again
# nightpool[m] `to: "https://cybre.space/@nightpool` works because it's just referencing the actor object, but `to:
{type: "Link", href: "https://cybre.space/@nightpool}` doesn't make any sense.# jdormit[m] That makes sense
# cjslep[m] I think the 'range' of the 'to' property in ActivityStreams is only descendants of "Object", which is explicitly disjoint from "Link" types and so it would be invalid to stuff a Link into a 'to' property.
# cjslep[m] Whereas JSON-LD permits using an IRI reference in place of any value, so addressing ` 'to': 'https://example.com/addison' ` is using a whole different concept than the Link ActivityStreams type.
# nightpool[m] yep
# nightpool[m] that's basically what I said
# cjslep[m] Yes, we're in agreement. I just wanted to be sure to emphasize the nuance of one being a JSON-LD structure concept (1 layer down) and the other an ActivityStreams Vocabulary choice (surface level semantics). Despite appearing to be very similar.
labs[m] joined the channel
# jdormit[m] cjslep: ah that makes sense, thank you for explaining!
vitalyster left the channel