Kinda off topic but it's a pity about the choice of F#, even though it might be a fantastic functional language. I just can't find myself willing to invest time learning a language that's basically locked to a single operating system.
Most of my work is on software that runs on big iron unix systems and I suspect Mono would be a tough to impossible sell in large enterprise due to lack of professional services and support. Hell it's impossible to even sell the idea of putting Linux on some of the systems unless it's RHEL.
Exactly what I was thinking. I know functional programming is interesting to me, and from experience I know that learning a new paradigm can improve your competence as a developer, but I have the same reservations about learning a language I'm pretty sure I won't have a realistic chance to actually use.
If there are similar resources in Haskell or Erlang... now that I could convince myself to spend some time on.
I've studied some ML and as far as I can tell, F# is basically just ML plus access to CLI libraries and support for CLI-specific data structures, generics, etc. Awesome if you're a .NET developer, but if you aren't, then there's not much of a point in learning F#--but SML and OCaml are basically equivalent options.
F#, SML, and OCaml are basically one functional language family, while Haskell and Erlang are in completely different families. And then there's the Lisp/Scheme family, which is again quite different.
Interesting! All I know about .NET is that the native ORM is apparently awesome (second-hand opinion, no less, and I can't even correctly recall its name).
I learned a little Scheme to try and help a friend while I was in college, but that's about the extent of my exposure to functional languages... it seems like a real shame that this style of development is apparently so highly-regarded but doesn't get much attention in early programmer education.
I definitely see the utility of OOP, and as far as marketability goes it's still the good bet, but it'd be great to see better resources for functional programming.
The ML family is quite different from Scheme. ML's salient feature is its powerful type checking system. It allows you to almost eliminate runtime errors.
Functional programming used to get much more attention from universities than it does now. So did C. Now they're trending toward doing everything in Java, which is what industry seems to want, but it's dumbing down the education.
Yep, that's the one! Thanks again. Despite often not totally agreeing with Joel, I taught myself and my classmates C and then C++ in high school, re-learned C++ in college, and am one of those people who railed against Java as a replacement.
My knee-jerk reaction would be that functional programming isn't an important skill for industry... but then neither are the low-level RDBMS classes I took as requirements, and I'm sure a lot of other people could chime in with "US university failed to prepare me for work" stories of their own, so that's beaten to death.
I'm getting over my reactive Java loathing thanks to Android, but man, it'd still be nice to get people learning functional and procedural and object-oriented programming.
What's fantastic about Java is not the Java language but that you now get a JVM for just about every platform. You have Java, Groovy, Scala, Closure, etc. all these programming languages you could pick from and you get cross-platform support for free as well as being able to use all the thousands of Java libraries that are already out there. It's a powerful combination.
Most of my work is on software that runs on big iron unix systems and I suspect Mono would be a tough to impossible sell in large enterprise due to lack of professional services and support. Hell it's impossible to even sell the idea of putting Linux on some of the systems unless it's RHEL.