Java is a turing-complete language. That means that anything that is computable across turing machines (computers, pretty much), can be computable in Java. However, this does *not* mean all turing complete languages are equally good at all tasks. Anyway, Scheme and Common Lisp are pretty good at this sort of thing.

Quote: However, this does *not* mean all turing complete languages are equally good at all tasks. Anyway, Scheme and Common Lisp are pretty good at this sort of thing.

So, in ur opinion, are you saying that I should or shouldnt use Java?

Use Java.

The quite considerable benefits of using a language that you have a lot of experience with outweigh the meagre benefits of using a language that is (and this next part is just my opinion) more powerful, once you know it very well.

Having said that, if you're going to be working on these things for the next five years (or more), then trying a variety of languages (at least one of which should be Lisp) would probably be to your advantage.

John B

Quote:
Also time and memory performance is very bad

Rubbish. I can almost guarantee that for any problem suited to Prolog the Prolog implementation will run faster than your ad hoc implementation in another language. This also leaves aside the fact that to solve the problem in another language would require implementing sizeable chunks of Prolog's inference engine.

If the OP is looking into Prolog, he may want to take a look at Visual Prolog, based off of Borland's Turbo Prolog. It's a compiled language with many libraries (including windowing libraries), a debugger, windows designer etc.

Quote: Original post by MDI
[...]This also leaves aside the fact that to solve the problem in another language would require implementing sizeable chunks of Prolog's inference engine.[...]

In some languages, though, that isn't too difficult.

I suggest trying in Java, and if you get stuck, reading up on other languages so you can get a feel for other potential approaches. If you find you have a hard time implementing any solution with in Java, and you see that one of the languages you've looked into provides features that would make implementing a solution easier, then think about switching languages.

There is no need to use the best tool for a job, and people get by fine using suboptimal tools all the time. When factors besides the tool itself are taken into account, such as familiarity with the tool, it becomes much more difficult to pick a single tool that is objectively better.