Lambda the Ultimate

Speaking of which, I'd like to see what you have to say about my comment on your warp 9.99 blog entry.

And for those following along at home: Achilleas' blog post

Thanks for the link, a very nice blog post.

I'd say that overall, I almost agree with the blog post... only because my C++ code is just a semi-compiled form of Haskell. I use ubiqutous referential transparency (lots of const) and often compiler-specific annotations to mark up intent (that a better type system would allow, with much more intuitive and precise semantic meaning). I tend to be very abstract data orientated -- in the sense that I tend to reify anything that needs processing. Side-effects are compartmentalised as much as possible. State, if it's ever used, is local (on the stack, as loop counters, etc.)

So yes. Most of the advantages of FP can be felt outside of an FP language. But why hurt yourself (unless someone is paying you to do it... and paying well) when you might as well just use the real thing? Maybe it really isn't much more than syntax (and some clever preprocessor macros), but damn it, I want to enjoy my life and my work -- and I wanna look pretty!

Incidentally, I "learnt" ML after getting to grips with Haskell. The comments from people who point out the elegance and syntactic beauty in ML always makes me giggle.

I have never seen anyone point out the syntactic beauty in ML.

I've admired the syntactic beauty of the ML code in Implementing sets efficiently in a functional language. That paper describes how to write a balanced-tree data structure very clearly. I used it as the basis for my cord data structure in Ermacs. (I still fire up Ermacs occasionally on files that are too big for Emacs to handle.)