Conversation
Notices
-
Embed this notice
in some ways web tech feels like they just took what greybeards did and what if we golf the specs down by just declaring as much behavior implementation defined as possible and hope for the best :blobcatglowsticks:
like okay you have protobuf instead of asn1, and you have [gesticulates] this massive stack of pseudo-HTTP instead of corba, but what you are doing with all of it is the same shit they were doing in the 60s but now instead of one 600 page book about how it all works its 40 mini drafts of equivalents of the same things
idk
- Blurry Moon likes this.
-
Embed this notice
@newt
> asn.1
i've read arguments against it and they are not very convincing. they mostly boil down to, in order
- gnu didn't make a free tool for it so i hate it
- the spec is too big (because its complete)
- it went through several drafts and got a bad reputation for being hard to learn before it wasn't
the last one is the only one i somewhat care about. but that -was- addressed later on, and there's books that further fix it since years.
> corba
i don't agree that it was worse. it seems like its about equivalent but aspired to describe the entire stack.
-
Embed this notice
@icedquinn are you trying to imply that ASN.1 was good design? Do I need to remind you that SNMP is described in about 30 different RFCs?
Whatever your grief with today’s webshit is, the things before it were certainly not better. In fact, they were much worse (like CORBA)
-
Embed this notice
@icedquinn much of practical business labor type computing is just people avoiding abstract thinking like the plague
and this makes perfect sense and is successful in a lot of contexts
-
Embed this notice
@whiteline something i noticed with those is they just went and did shit because they had a real need for it.
somebody comes behind to package up what people actually did. this becomes some new package or report.
the common lisp design model (bottom up) works extremely well and its how we should do everything :blobcatgoogly: