I do have a remark though, which is about OCAML (and JOCAML). Although the syntax is a bit awkward, the same holds for any ‘species’ of programming language you’ll learn. I remember that my first Basic (in the early 80′s) experience included getting used to the strange syntax, just as my first assembly, prolog, lisp, C/C++/Java, Haskell, Python and OCAML experience. Just because someone is used to a certain syntax doesn’t mean that other syntax can’t be good either (I prefer Python and Haskell, although that also take time to get used to). After a while most ‘weird’ syntax issues disappear once a person starts to think in the logic of the language, and not the syntax (although I never got used to LISP, but that is something VERY personal, and not a reason not to recommend it).

OCAML has a really good performance (close to C) but is high level. It can be used for functional and imperative programming and has classes.

The reason for the ‘+’ and the ‘+.’ is because it is so strongly typed that the compiler has different optimizations for integers and real numbers, which after thinking about it makes sense. Integers are often only used for counting (in arrays etc), and reals for the more advanced mathematics.

Please have another look at OCAML. Especially with the JOCAML it is muticore functional programming at its best. No wonder that Microsoft based F# on OCAML.

Having taken the time to learn Scala myself, I’m pretty happy with it. Its also nice seeing more momentum with its uptake and the number of open source projects starting that are implemented in Scala.

I still think its too early to tell if it will unseat Java in terms of new (enterprise) project numbers (its nowhere near that point now). I think much of that will depend on when Java 7 is released and reaction to it.

To me, the advantage is that you don’t end up with insane interface pollution…. i.e., interfaces everywhere, by the hundreds, one for every method that wants to accept multiple different kinds of objects.

In Ruby, you can print any object that has a .to_s method. You don’t have to create a Stringable interface, and all those objects don’t have to inherit from String. Just implement .to_s and throw it in there.

To me, that’s the big win.

While I agree with Scala as a winner, it is not the easiest language to learn. Compared to for example Python, it will take a bit longer time to do much useful.

Interesting analysis, thanks. I agree with the second commenter: not only should have Fan been in your list but I think it would end up in the top two with Scala (and personally, I would pick Fan over Scala).

– Cedric

I have to say, the comments degraded into the sort of dick-swinging, alpha-nerd, overtly defensive, pointless one-upmanship that makes me want to abandon software development.

The object syntax is too verbose but it’s not a deal breaker (and it’s pretty transparent what goes on underneath so once you get it, you get it). If you write new code you can choose a more modern syntax. That’s the reason Perl 6 is still vaporware, by the way, Perl 5 is still in active development.

It’s simple, it’s pretty fast for an interpreted language (faster than Python and Ruby), it’s got pretty solid Unicode support (hardly any other has, except perhaps Python), and the massive amount of free code at CPAN is a life-saver.

It still may not be the #1 language to learn if you should learn only one; In my opinion that would be C due to it’s universality and being only halfway-high-level so any structure and technique you learn is universally useful. But you count it out for all the wrong reasons.