Sure MatLab is the way to go. It helps a lot when there is no need to test algorithms with huge amounts of data, and when these algorithms could be refined from the problem area.
I guess that ideally one uses some built-in scripting language, to implement higher-level algorithms, and these that are of the lower level - still have to be written in C++ or something like.

Halloween, I''m not entirely sure if your for or against Matlab!?

For those that are interested...

The beauty of Matlab is that like Java, it''s an interpreted language. Basically it''s a form of C script, so if you want to make a small change as Ferretman suggests, you are not wasting time recompiling. Much of the power of C is hidden within built in Matlab library functions so, if you want/need to do something very low level, then yes you can write your own code to do that.
There is no deteriment to using Matlab with large amounts of data as Halloween seems to suggest. Certainly, being an interpreted language, Matlab runs slower than native C. However, the power of Matlab with data is its various forms of the array class. Take for example a simple matrix multiplication. In C or most other languages, you need to write out the multiplication element by element (and hope you dont get an index wrong). In Matlab, you just use the array names in one multiplication equation.

Matlabs real power is in data manipulation and visualisation. There are other scripting languages that have more visualisation than Matlab (IDL for instance) but Matlab is far easier to learn. you can even make games in Matlab (although they dont run very fast!).

I use Matlab predominantly because the sort of AI development I do involves large amounts of data and/or complex mathematical operations. It''s far easier to do it at a high level than to try and implement it in C/C++!

Cheers,

Timkin

There are still other languages around besides C++?

I couldn''t figure out how to get to the last post in this thread. Any Help?

IceStorm