+bonifartius 𒂼𒄄
AP rant
activity pub is not only what happens when you let OOP people design a protocol. no the OOP try as hard as possible to make things as complicated as possible. kinda like the ITU standards, X.500 n stuff. utterly deranged.
i did ldap related things for a while which is a reduced version of X.500 and that gave me the same vibes as AP. everything reinvented from basic principles. beautiful contraptions with their own lingo and everything.
it's like if you have a box full of ultra special lego parts. they are beautiful in their right place, but without their set and the manual they are absolutely shit.
everyone does things a tad different though - the protocol is so complicated, it can't really be implemented.
to add insult to injury, the "normative" thing isn't a machine readable version, but a fucking giant free form web page. it honestly is a wonder that there is more than one implementation, and that those can speak to each other.
AP rant
@niclas i really don't get why this is done time after time. W3 standards are doing this, freedesktop.org does this, etc.
meanwhile the still relevant standards are one-file rfcs from like 40 years ago :)
AP rant
@bonifartius I know (via Apache Software Foundation) Roy T Fielding, and he wrote https://www.rfc-editor.org/rfc/rfc1945 basically by himself.
I met him a couple of times, and he told the story; "Oh, this web thing is cool... 35 sites and I read it all. A month later there was 200 sites. This is getting awfully popular, I think we need a formal specification."
He must be one of a very small number of people who has at one time "Read ALL content on the world wide web"
AP rant
@niclas haha :D
http1 was sane. http 2 and 3 are just reinventions of things other layers and protocols should do. mostly because NAT and firewalls blocking everything but 80 & 443 ..
it's really dire if you think about it
AP rant
@bonifartius Overly complex standards are typically a sign of "design by committee" where everyone has their own pet peeve that "must" be in there.