← Back to context

Comment by 082349872349872

3 months ago

for "Semi-Dynamic Language" it might be worth looking into rpython: interpreters written in rpython have two phases, in the first phase one has full python semantics, but in the second phase everything is assumed to be less dynamic, more restricted (the r of rpython?) so the residual interpreter is then transpiled to C sources, which, although compiled, can also make use of the built-in GC and JIT.

8 comments

082349872349872

Reply
MillironX  3 months ago

I immediately thought of Julia as a semi-dynamic language. Julia is a dynamic language, but (as I understand it) the first time a function is called with a specific type signature, that specific method is JIT compiled as static LLVM.

  • Jtsummers  3 months ago

    Which is then used for future dispatches on that same signature and gives it very good performance. Julia is dynamic, and definitely beats the 10x slower than C barrier jerf mentioned.

    For what I was using it for at the time (~3 years ago when I used it seriously) it offered performance close to the compiled orbital analysis code we had (in more conventional languages, Fortran and C) but with the flexibility for developing models of Python and other dynamic/interactive languages. An excellent tradeoff: very small performance cost for better interactivity and flexibility.

gsf_emergency  3 months ago

Of all the non-esolangs (=exolangs?) APL+kith seem to be almost designorismic/estuarine (formerly, Riverian) beasts..

In your informed opinion, how would it make sense to be thrilled thinking about (not just a semi-dynamic APL (=S-DAPL?) as above but) designing

  a APL-kith for widespread adoption by BOTH (demoscenic)gamecoders & Analysts (the academic variety)???

Specifically, which of the sneering checklist items[0] would be killer to cross off?

[0] https://www.mcmillen.dev/language_checklist.html

(Note in particular that the very APL inspired Wolfram has monopoly with physicists BUT does nothing for engineers

https://www.stephenwolfram.com/media/physics-whiz-goes-into-...

1988

>“But we tricked him, so to speak,” says Nobelist Murray Gell-Mann, who helped to bring Wolfram west. “We gave him a Ph.D.” )

  • 082349872349872  3 months ago

    Well, I'm thrilled by something in the neighbourhood, but your widespread adoption there is more a "well known in small circles" kind of thing: instead of selling shovels to the gold miners, the target of demosceners+analysts sounds like selling paint brushes to the garret-dwelling artists?

    For me, I guess the killer checklist item is "[X] Programming in this language is an adequate punishment for inventing it".

    For this putative APL-kith, I'd guess the combination of "[X] You require the compiler to be present at runtime [X] You require the language runtime to be present at compile-time" would be killer; after all wirth-style compilers, these days, can run out of L1$. However, does sufficient staging run into the "fewer than 100 programmers un-algoled* enough" problem?

    (could relative popularity of the square bracket language be because phykers display formulae in order to convey insights, but for imkers the calculations are their own point?)

    * on the "[X] Rejection of orthodox systems programming without justification" front, what might be to algol-inspired programming as Thistlethwaite's algorithm is to plain old cube solving? https://news.ycombinator.com/item?id=42426716 instead of reducing the group operators as one proceeds, one would reduce the dynamism as the loops nested... (one of the things I find impressive about the rpython JIT is that it roughly manages to implicitly do that reduction!)