#social 2017-10-13

2017-10-13 UTC
rowan, rowan_, timbl, cdchapman and xmpp-social joined the channel
#
cwebber nightpool: Gargron: rowan isn't on irc at the moment, but I wonder if you could look at their issue? they're trying to get interop with Mastodon, and are looking at the webfinger side of things, which I don't know much about https://mastodon.social/@rowanlupton/19164612
#
Loqi [rowanlupton] @cwebber in webfinger/activitypub, how does masto make the leap from webfinger's response to finding the user's feed? i figure i must be missing something, because @roo doesn't show up any posts, when i think it should?
timbl and rowan joined the channel
#
cwebber hi rowan !
#
rowan hey hey!
#
rowan what's up?
#
cwebber rowan: just saying hello... sorry I didn't give you a clearer response on your webmention question
#
cwebber it's a bit out of my area
#
rowan no prob
#
rowan i don't think that was me
#
cwebber oh
#
cwebber https://mastodon.social/@rowanlupton/19164612 isn't you?
#
Loqi [rowanlupton] @cwebber in webfinger/activitypub, how does masto make the leap from webfinger's response to finding the user's feed? i figure i must be missing something, because @roo doesn't show up any posts, when i think it should?
#
cwebber oh hey nightpool responded :)
#
rowan oh it was me a long time ago :)
#
rowan i was like "i haven't been on masto or here in two weeks"
#
cwebber oh wow
#
cwebber why did it show up 13 hours ago on my timeline
#
rowan hahaha
#
rowan i have no idea!
#
Loqi nice
rowan, JanKusanagi, jankusanagi_, cdchapman and surinna joined the channel
#
ajordan cwebber (or anyone else really): do you know of a good canonical "why .well-known is problematic" explanation?
#
cwebber ajordan: "you can't use it with two applications on the same domain at once"
#
cwebber there's one single-sentence answer
#
ajordan oh is that it? I thought there was more
#
cwebber there are more reasons :)
#
cwebber ajordan: the converse isn't just "why .well-known is problematic", but also "why follow your nose is good"
#
cwebber think of how the web works: it's effectively a graph of links
#
cwebber you can jump from one to another to another
#
cwebber now consider that this breaks once you rely on .well-known to have "per-server" resources
#
ajordan ah yeah
#
ajordan makes sense, thanks
#
cwebber the other thing is
#
cwebber right now we have the vision of the web as kind of being large-nodes
#
cwebber or even medium-nodes
#
cwebber imagine a future where a node going down doesn't hurt a bunch of people at once :)
#
cwebber well that future probably isn't compatible with .well-known
#
cwebber or, the rationale for it
#
cwebber doesn't make *lots* of sense
#
ajordan context, btw: https://github.com/securitytxt/security-txt
#
ajordan gonna file an issue but want to be eloquent :P
#
Loqi [securitytxt] security-txt: A "standard" that allows websites to define security policies.
jankusanagi_ and timbl joined the channel
#
saranix "you can't use it with two applications on the same domain at once" -- the fact is, if .well-known is really worth doing (it isn't), it would be worth putting into the HTTP protocol where it logically already exists instead of some url pseudo-namespace. TLS isn't started by grabbing .well-known/site-certificate, which is logically the same thing
#
saranix too many modern stacks are like 7 new layers on top of OSI. It's insane.
jaywink joined the channel