Reading this, I wonder how it measures up to Julia. It is my understanding that Julia was implemented in Lisp and so this almost reads like "we dropped the JIT thing and just compile instead"
For hard-core numerical computing, Julia would be, I think, be far superior.
Speaking as a 40+ year Common Lisp developer, I like this project because it is built on Common Lisp arrays and smoothly integrates with the Common Lisp ecosystem. I have spent some time using alternative CL libraries that sit on top of BLAS, etc.
> It is my understanding that Julia was implemented in Lisp
That's mostly incorrect. Julia's parsing and lowering are implemented in Femtolisp (though active work is underway to replace the parser with one written in julia itself). There's no lisp code doing any actual important or performance sensitive work in Julia.
The very small core of the language implementation is done in C and (a restricted subset of) C++, and then the rest of the language is boostrapped in Julia itself.
4 comments
[ 2.0 ms ] story [ 21.3 ms ] threadIn what way?
Julia is of course an entire language ecosystem whereas lisp-stat is a collection of packages mostly maintained by a single dude.
Speaking as a 40+ year Common Lisp developer, I like this project because it is built on Common Lisp arrays and smoothly integrates with the Common Lisp ecosystem. I have spent some time using alternative CL libraries that sit on top of BLAS, etc.
That's mostly incorrect. Julia's parsing and lowering are implemented in Femtolisp (though active work is underway to replace the parser with one written in julia itself). There's no lisp code doing any actual important or performance sensitive work in Julia.
The very small core of the language implementation is done in C and (a restricted subset of) C++, and then the rest of the language is boostrapped in Julia itself.