I haven't posted much related to programming lately. Looks like I've gone about as far as I can with Clojure, given the limitations of the JVM. It's really a shame because I'd love to use it more. There's no way around it, Java is simply not suitable for numerical computing due to the lack of libraries (for what I do, not necessarily what anyone else does) and the speed difference between Java and native for a handful of tasks like linear algebra. I've done a lot of research on it and concluded that Java will always be a good deal slower for some things (jblas may be a step in the right direction).
Maybe clojure-py is the answer. It offers a native Clojure solution. For now, due to time constraints, I've decided to set Clojure aside. This will most likely be the end of blogging about my Clojure and Lisp experiences. It's been a great journey and I hope that in the future I will
be able to program in Lisp on a regular basis. I may occasionally comment on a piece of news in the Lisp world, but that's it.
Where will I go? The honest answer is that my current combination of R and Fortran works well. R is not a bad language if, like me, you believe in functional programming.
For a while, Julia was getting a lot of attention. It has potential. In a year or two it might be worthwhile to dig into it. It has a lot of features, but there's not a lot that's new, so I have no motivation to work with it now.
Haskell and OCaml would be great languages to learn for the sake of learning, but given my time constraints right now, I'm not sure the benefits would outweight the costs. (F# is not going to happen either.) I don't see either as realistic replacements for R.
How about Scheme, Common Lisp, or newLISP? Each has its uses, and I still use newLISP, but they also have drawbacks. Scheme documentation is crap, plain and simple. I know, Racket has good documentation of the language, blah blah blah, I've tried Scheme and I'm not smart enough for it. Common Lisp could be great, but a language that doesn't evolve is useless. I'll use the next revision of the Common Lisp standard (LOL) that has been updated for the 21st Century. newLISP is a scripting language, and I love using it in that capacity, but it will take more than that to replace R.
I'm not interested in Go. It's a better C. I have as much use for a better C as I do for a better Mandarin. If you don't speak Mandarin, and don't have a reason to speak Mandarin, a better Mandarin is not going to excite you.
I may give D another try. I like the language. It provides little in the way of numerical libraries so my best guess is that I would be better off with Clojure than with D.
That's where things sit. When I have something worth blogging about, I will do so.