pouët.net

What so you think about the "new" D programming language? IMHO this may be the new language hype. Read the Digital Mars page, the concept looks really cool.
Flame on fellow poueteers :)

anything used to code this beauty cant possibly be any bad

Argh, I read that as "Parse C 47" -_-

i've been watching D pretty closely for about 3 (or 4?) years now. i'm very very interested, though i've yet to start a big project using it :)

D has very nice advantages compared to the other modern OOP languages. But if only you are an advanced coder. If you are not, use Pure Basic instead :)

"Using D Header files at DedicateD for OpenGL and SDL and at D - porting for SDL_mixer."

Programming humor. Funny, haha.

I used to be fond of D, but I don't program much at the moment.

Not, at least, until there are official headers for Direct3D from MS.

Let's wait to see which apis will be supported in D (opengl, d3d). At the moment the best choice is still C# and managed DX.

imbusy: D doesn't need D3D.. stop being such an M$ fanboy. D3D itself is no reason to not try D.. d3d is not _that_ great.. also D with openGL will most likely soon run on the mac in addition to linux.

in comparison: with C++ (until microsoft bastardized it to keep people from writing multi-platform code) the environment around it was a platform-independent/didn't promote any specific platform. D should stay that way, I.E. it's better if microsoft don't release anything because then their fanboys will start using their "extensions" or "improvements" and whatnot and thus hurt the concept of the language and its environment. M$ won't do anything "good" if it doesn't mean they can bind more fanboy coders to their monocultural platform.

(M$ C++ is incompatible with the rest of the C++ world which runs on linux/mac/bsd/whatever)

these arguments are not true for demos in particular who tend to be very platform bound, but in a general perspective.

i currently am using D for several projects of mine.

jaw: msvc8 c++ is more compatible to language specifications than gcc or g++ (which still don't support local classes), though it's not still perfect.

but yes i can't even express myself in msvc6 without hitting my head on the wall.

Jaw: you're really impressing us with your 1997-style "M$ W!ind0z3 sux!"-style arguments.

Myself, I think D looks very promising, but I do feel that they haven't properly addressed some of the current "big issues" with programming languages - multi-processing. For something to "replace" C/C++ today, I think parallelism must be seriously worked on.

c++ will feature parallel execution paths in the future according to here.

This is only a small step. Look at languages like Haskell or Fortress for more complete approaches.

But they encompass a completely different concept of programming.

What kusma said. D is nothing more than a fix to some disadvantages of C++, and hardly spectacular in itself. And C++ is from the seventies, and it's 2007 now. Computers have become all fast and stuff, and we still fuck around with increasing pointers and strongly typed iterator classes. How about we tell the computer what to do instead, and let it figure out the "how" part itself?

Personally though, I prefer things like ruby over haskell and friends, but I guess that's just because that's the programming style I'm used to, and because it's well suited to the programming I do.

The slow speed of interpreted languages is often a lot less expensive than the increased development time, these days. I guess not all of this totally holds for democoding, but in general I'm just baffled by how much a thing like C++ is still an industry standard.

skrebbel, would you rather increase the development time 2 times rather than buying 10 times more the equipment in performance critical applications?

imbusy: If a developer cost ~50K eur a year, and new equipment is a one-time cost of the same, it's easy to prefer buying equipment over developers.

skrebbel: Yeay, let's use OSDM instead! ;-)

@skate: I don't know, but when I first learned C++ (and later C) I recall wishing that I could use D, since it allows for returning multiple variables from a function without having to recall whether you're referencing or dereferencing a pointer, or whether it's necessary or not because you have an array, etc etc. I think some of us largely resorted to trial and error at the beginning- keep switching the & and * around until it works somehow.

Seriously, that one feature would have made it worth my while.

Then again, I've become aware of far more practical uses for pointers (referencing variable-length items in lists, or other data types) where they're far more useful than the alternatives, but I never really got too far into that.

I think there is really a lot of new developer-friendly stuff in it. It has a better/safer concept than Java and might be still as fast as C++. Its true that it lacks parallel execution support though...
I really liked Java from a developers point of view, but it is too fucking slow...
Is D supported by any popular IDEs yet?

I've been using PowerD for amiga stuff...

rarefluid: they do say sun has been investing into making java faster, that it'll start showing on 5.. but somehow i dont believe one word of it until i see any of it working. if it becomes a reality all the new media art folks using processing out lack of patience for coding in a manly fashion (thats a private pun btw) would sure be pleased.

Haskell for the win!