I get it for the FORTRAN -> Java transition; I understand that you'd want to port your stuff to Python from Java. But why would anyone port massive and complex libraries (SciPy is not your regular "Yep, coded that in 4 days!"-gem) to a virtually identical language? [1]
No point. Learn Python, and get some stuff done.
[1] Now, if you're about to build a good scientific/numerical/statistical/(...) library for Clojure -- I'm all ears.
But where's the package management? Clojure might be a bit too young.
1. Use Maven as Maven with https://github.com/talios/clojure-maven-plugin - My impression is that this option is often more popular with people coming to clojure from java, as you get the full XML and maven experience.
2. Use Leiningen https://github.com/technomancy/leiningen – this is a project and dependancy management / built system built in clojure on top of Maven's libs. You have a single (and simple - no XML) project.clj file that describes your project. My impression here is that lein is popular with people coming to clojure from non-JVM land. Being a former pythonista, this is my preference.
There is also a default maven repo that i believe both work with hosted at http://clojars.org.
In both cases maven handles dependency resolution and package management. My experience is that a global install of jars is not common like it is in python or ruby land.
The other tool that you will want to look at is https://github.com/liebke/cljr which is a repl built ontop of lein's package management commands. This is useful because it lets you run a repl without a project (lein requires a project) but still gives you control over what libraries are available. If you are using incanter (the stats package mentioned by Estragon) you will want cljr
I'll have to check out cljr.
...if you're about to build a good scientific/numerical/statistical/(...)
library for Clojure -- I'm all ears.
Presumably you're aware of Incanter (http://incanter.org/), which wraps scientific libraries for Java.I'd love to see more people pitch in and help develop tools that would make scientific analysis easier.
I did a presentation at RubyKaigi on this general subject actually: http://www.slideshare.net/knowtheory/mapping-the-world-with-...
(and a video: http://rubykaigi.tdiary.net/20100829.html#p09 )