chris martens14h ago

re: https://hci.social/@chrisamaphone/116325060049701188

taking off my "impartial observer" hat for a moment (and so breaking out of the thread), one opinion i've started solidifying is that we need stable (or one might say "archival") proof languages, alongside those that actively evolve. a big motivation for me is to develop teaching materials that still run in a decade (Explaining), but i think there are good Convincing-aligned reasons to want this as well.

deech12h ago
Show threadchris martens
arguably we should also want archival programming languages more generally. sometimes, a language whose features cease to evolve is called "dead". but perhaps we should reserve "dead" for languages whose programs no longer run, and use "archival" for those whose implementations are maintained while their feature sets remain stable (thx @simrob for planting this lexical seed in my head)
6:29pmWeb
Show threadDevine Lu Linvega13h ago
@chrisamaphone Have you ever come across Kragen's notes on using UVMs for archival? This might be a good entry into the topic and probably has some usable ideas for your work:
https://dercuano.github.io/topics/archival.html
Archival ⁂ Dercuano

Show threadchris martens9h ago
@neauoire oooh i have not, looks very relevant! thanks!
Show threadKirtai 🏳️‍⚧️13h ago
@chrisamaphone @simrob
Languages with decent standards that implementations actually follow?
Show thread✧✦Catherine✦✧13h ago
@chrisamaphone @simrob I feel like maybe the ideal way of achieving this is compiling any existing language down to Wasm and archiving that
Show threadKirtai 🏳️‍⚧️10h ago

@whitequark @chrisamaphone @simrob
But which version of wasm? Assuming it doesn't become a "living standard".

I hate that term so much.

Show threadchris martens9h ago
@simrob @whitequark that is indeed a key piece of how certain examples i am thinking of currently live on (specifically via mlton-wasm)
Show threadchris martens9h ago
@simrob @whitequark it would be nice if we could count on wasm to be stable!
Show threadJoe Groff (1M Context)3h ago
@chrisamaphone @simrob @whitequark I wonder which will remain more stable over time, wasm or x86 win32
Show thread✧✦Catherine✦✧1h ago
@joe @chrisamaphone @simrob I don't think wasm will ever have a v2 so probably that (seeing as x86 win32 does in practice drift quite a bit)
Show threadMx. Chara Aznable (they/them) of Pnictogen12h ago
@chrisamaphone @simrob honestly they should never have called programmers' gabble by the honorable word "language", because that one misusage of the word really seems to be cause a lot of trouble. is an unchanging engineering specification "dead"? or is it simply not the sort of thing that should be changing? programming "language" (i.e. not real living language, but an artificial script for issuing commands to an inanimate device) arguably shouldn't be infinitely and eternally mutable. they're treated that way right now but I suspect the actual reason for this is a cynical one: a perpetually chaotic and undefined programming landscape is especially easy for grifters—I mean, "venture capitalists" and "entrepreneurs"—to monetize. Profiteering computer geeks can pose as saviours forever in such a dysfunctional landscape, always claiming to have sure-fire (or at least salable) routes through the ceaselessly shifting chaos.
Show threaddeech12h ago
@chrisamaphone @simrob In the case of non-mainstream languages for it to be archival there needs to be a reliable bootstrapping story that starts with some mainstream language (edit: ideally with multiple implementations).
Show threadConor Titania Mc Bride12h ago
@chrisamaphone The diachronic life of programming languages is tragically neglected, but a vital concern. @simrob
Show threadHeath Borders3h ago
@chrisamaphone @simrob iOS developers felt that during Swift's first 5-7 years as the language evolved and broke past programs while Objective-C kept chugging along working as well as ever