About Julia and Lisp

Even if we could call FemtoLisp code, implementing scmutils is no breezy task. This port to chez-scheme is 71,000 lines of code. Given that femtolisp isn’t even a scheme implementation, I think that’d be a gigantic undertaking to port to femtolisp.

2 Likes

I was not suggesting to port scmutils to FemtoLisp.

Also, traditionally the definition of “Scheme implementation” is rather wide, and on the FemtoLisp page one reads things such as:

highly compatible with Scheme, including some R6RS features

it is fast, ranking among the fastest non-native-compiled Scheme implementations

femtolisp is a simple, elegant Scheme dialect.

It seems that the below issue is more pertinent to what I had in mind, and indeed cuts it short for the time being.

It also works with PSL version (that’s what it was based on originally, however I deprecated that in favor for CSL since it supports unicode characters also. The REDUCE developers do not have a type system in their language, unlike Julia. So I believe that ultimately, the best solution for symbolic computation with Julia would be to port something like REDUCE into the native Julia, since they complement each other and each are missing the features the other offers.

I would certainly be happy to work on the full implementation, if funding for such a project exists… in fact I have already communicated with the REDUCE developers about how it can be done.

Unless someone funds that though, I am happy with how I developed the Reduce package as a demonstrator of what I would like to see fully integrated into the Julia language someday.

2 Likes