anybody read the concept by Dr. Alex Nareyek called Excalibur? Some example implementation made by him and his team is Dragonbreath -> find more info at www.ai-center.com - it's quite interesting reading... the question : do you think, that when this concept is properly written in c++, that it will be fast enough to compute, say 10 - 20 entities in some action game? because the concept is based on graph isomorphism and this is not a very trivial task thanks

hmm, why'd you post this in Everything Unix?

wouldn't the AI forum be a better place?

Why would the language matter that much? It may change the performance some, but the algorithm is what's really important.

this is my first post so I put it in wrong place, sorry. I hope I can manage it next time :-)

- I know the algorithm is most important, but I saw the demo of the technology (www.ai-center.com) and I am not pretty sure about garbage collecting in such a code where the allocations are used quite often, and such allocation is not leading to solution most of the time (searching for structure of the problem) so the memory is wasted (for some time until garbage collector does not free it).
anyway - the demo technology is just something like a proof - that it can be done this way - and it looks like it works, but the performance is not the goal of the demo. that's why I am asking about it. the concept's fine, but will it be fast enough when the concept is rewritten with the performance in mind?

thanks