#social 2019-05-16
2019-05-16 UTC
DenSchub, vitalyster, xmpp-social, Guest84, bblfish, Simounet, bblfish_, saranix, timbl and schmittlauch[m] joined the channel; vitalyster and remmalj[m] left the channel
# 
schmittlauch[m] Question regarding post IDs: The AP spec always talks about HTTPS URIs. But shouldn't that be URLs, as the post has to be retrievable form there -> its *location* has to be given, not just its unique identity?
# rialtate[m] I thought the spec always says IRI
# schmittlauch[m] @rialtate:matrix.org Yes, it says IRI. But also "Every URI [RFC3986] is also an IRI, so a URI may be used wherever an IRI is named"
# schmittlauch[m] and we're still at the "identifier" (without addressing information) vs "locator" (with addressing information for retrieval)
timbl joined the channel
# rialtate[m] Who here remembers when tv people used to say "dub dub dub" for web addresses?
# schmittlauch[m] So am I assuming wrongly that posts shall always be retrievable via their ID (not taking into account different visibilities) and it's just the current way people are implmenting it, but it could be handled differently as well?
timbl joined the channel
# rialtate[m] "posts shall always be retrievable via their ID (not taking into account different visibilities)"
# rialtate[m] Pretty sure the spec says almost exactly that
# schmittlauch[m] That's why I'm confused that authors chose IRI/ URI, not a locator
# rialtate[m] Maybe try pinging @cwebber@octodon.social (co chair of spec)
# fr33domlover schmittlauch[m], iirc there's a "url" property in AS2, I don't remember its description but take a look
# melody data URIs are arguably dereferenceable right? or do those count as locators?
# cjslep[m] I think IRIs are not bound to the http protocol (unlike URLs) so theoretically part of the linked data graph could be content addressable instead of location based addressing
# schmittlauch[m] That might be a valid point. I gues I'll ask cwebbe, he loves content addressable stuff.
ajordan joined the channel