Re:Quilc: An optimizing quantum compiler written in Common Lisp
by aidenn0 on Thursday February 28, 2019 @08:18PM
(#19277148)
The macro system is primitive in the sense that a hammer is primitive: old but still a good choice for putting a nail in something.
I don't want to get into a scheme/lisp flamewar (there are enough of those already), but I think the expressiveness of the two is sufficiently close to make other choices dominate[1].
As far as OCaml, despite what appears to be mutual love between the ML and lisp communities, the world can be divided into people who like typed languages and people who like untyped languages, so I imagine that has an effect.
Ultimately there's a lot that goes into making decisions into what language to use; I think the Mill guys are insane for implementing their assembler in the C++ templating language (as in a valid assembly file is sugared into a C++ program that outputs the binary), but it works for them, so shrug.
1: OTOH Scheme is small enough that various implementations are doing intereseting things with expressiveness (racket has several, the language modules coming to mind first; Gerbil with process migration, &c.). Common Lisp is probably just too big to make a new implementation just to experiment with a single feature (though see CLASP for a counterexample)
I don't want to get into a scheme/lisp flamewar (there are enough of those already), but I think the expressiveness of the two is sufficiently close to make other choices dominate[1].
As far as OCaml, despite what appears to be mutual love between the ML and lisp communities, the world can be divided into people who like typed languages and people who like untyped languages, so I imagine that has an effect.
Ultimately there's a lot that goes into making decisions into what language to use; I think the Mill guys are insane for implementing their assembler in the C++ templating language (as in a valid assembly file is sugared into a C++ program that outputs the binary), but it works for them, so shrug.
1: OTOH Scheme is small enough that various implementations are doing intereseting things with expressiveness (racket has several, the language modules coming to mind first; Gerbil with process migration, &c.). Common Lisp is probably just too big to make a new implementation just to experiment with a single feature (though see CLASP for a counterexample)