Poll #29: 06/08/02 - 07/09/02



AI SDK/Middleware Focus
What area or technology should an AI SDK (middleware) focus on to get you interested?
[120 votes total]

Providing good, fast pathfinding libraries my units can call. (11) 9%
Providing good, fast FSM and FuSM libraries from which I can craft an AI engine. (12) 10%
Providing a robust scripting engine which I can use to build an AI and insert it into the code. (16) 13 %
Providing pre-built decision trees for common operations such as attacking and healing decision making. (2) 2%
Providing pre-built modules to handle basic and common "tech tree" decision making. (1) 1%
Providing libraries to quickly and easily set up distinct personalities for my game agents. (4) 3%
Providing some mechanism by which my game agents could quickly learn from prior experience. (13) 11%
Provide basic neural network and genetic algorithm libraries. (16) 13%
It's got to provide a whole, ready-to-roll game AI engine, or it's just not worth it to me. (15) 13%
Middleware? I don't need no stink'in middleware! (30) 25%


Steve's Comments: Interesting, very interesting. While a strong percentage don't think any kind of AI SDK or middleware would be of use, most folks seem interested in the possibilities if it truly has something unique to offer. Pathfinding doesn't seem to be a big draw (probably because most developers already feel they've got a handle on that), but no one technology seems to be an obvious focus either. That makes things tough for the potential AI middleware folks out there, since they don't have a lot of areas to concentrate on--they'll be rather more hit and miss and focused on individual large users and their needs for a given, high profile project than they probably would like. Oh well...it'll be interesting to see how this all develops.