I dunno, seems like exactly the type of situation where it would be much more appropriate to just write the damn thing in Java...isn't the ability to do just th

I dunno, seems like exactly the type of situation where it would be much more appropriate to just write the damn thing in Java...isn't the ability to do just that with performance sensitive code one of the best things Clojure offers compared to "standard" Lisps? You'd get better performance, clearer (and less) code, and less development time, as well as the ability to practically cut and paste the C code from the linked paper. I feel (as I do so often when projects written in certain languages that may or may not start with "Haskell" show up here) that the main take away is "Look, we can write working apps, too, and they're Better-Because-They're-Functional!" And yes, there are a lot of things that are better when you code them functionally, but simple numerical algorithms like this one play specifically to the strengths of imperative languages, and I see no clear benefit. It's the same code smell that I get when people start putting together utility libraries, macro tools, and bytecode processors just so they can pretend to do functional programming in Java: if you find yourself struggling to do things non-idiomatically in one language that would be trivially natural in another, why not just switch languages? Especially if they all live on the same JVM and play fairly well together...