I doubt that's all of my reasons, but not all of them are conscious choices or easily put into words. Higher or lower level control, more convenient paradigms for language construction (especially relevant since it's often easier to implement languages in a language of the same paradigm), etc. The (much easier) potential to later bootstrap my language (implement it in itself, using a previous version of the compiler to compile newer versions). Here's some reasons I choose to use other languages instead:Įnjoying the process and fine gained control of doing everything myself instead of relying on someone else's work to found myself on (this varies by degrees, you'll always have someone else's work in the process at some point, but you get my point). Not wanting to tie myself to Racket's language, execution model, ecosystem, tooling, community. Not having experience with or desire to learn the Racket language, execution model, ecosystem, tooling, community. Not liking lisps as production languages (cool in concept, clunky in practice). If you want the reasons behind the reason, though, that's a little more complicated - it comes down to a combination of a few things for me: It's not a particularly deep question, if you actually think about it. That's the only reason I or anyone else really needs.
#Racket programming language code#
Libraries and code written for racket may be right for racket, but are they right to me? Do they fit together correctly? I also care a great deal about conceptual integrity (as identified by fred brooks). Risc-v has been criticised for aligning itself overly closely with c semantics. This blog post complains about the state of llvm's gc primitives, and I believe gccgo frequently makes slower code than the reference go compiler because of the gc. More recently, wasm designers did the same thing to lisp implementors. net designers sought advice from apl implementors, and then ignored it because it did not suit them. This is a general problem that prospectively 'universal' runtimes run into. Performance characteristics of scheme are not universally applicable. Racket is, as you say, a platform it has already such a model, and it is not the same as mine. My interest is not simply in specifying a set of semantics but a model for interaction between humans and computers. I am not convinced of that, but it is something to think about. Dan Ingalls says an operating system is a collection of things that do not fit in a programming language. Many PL people do not care about operating systems.
0 kommentar(er)
