First of all, what are some good rules for an ALife system. Go towards friends. Fight or flight with enemies. any others? By ''good'', I mean rules that will lead to complex yet simple and beautiful patterns. I have seen Boids and Floys and I like the way those work. Now, I am looking for other rules that could be used. Second, how do you implement ALife. For example, how do you keep track of what each individual knows about other individuals (i.e. their positions, which are important for flocking)? Do you just give them global knowledge of all individuals positions? How do you store this data? Think outside the dodecahedron

perhaps each entity is reaction based, it doesnt know anything about anyone, it only knows about itself. It reacts based on it''s environment.

The rules are very dependent on what type of system you are trying to create. In boids there are 3 rules for each bird which are:
1. Try to get to the center of the flock
2. Try to maintain the same speed as the flock
3. Stay a certin distance from all birds in the flock

With these rules simulation can create a system of entities flocking, or schooling if you are modeling fish.

If you are just trying to create a life like organisms on the screen what you should do give each species a name. And then create a 2 dimensional array for each creature. Each time the species bumps into another species, it scores it for a reaction. If it is aggressive then you place a higher score. So it would be like the spcies bumps into Species 1, and gets an agressive response. So percentage of bumping againg =5%
[Species 1][5 percent]
and then the next time the species sees species 1, there will be a 5% chance that it will move away instead of investigating.

Get it? Or am I babbling. I created a Darwinism learning mechanism for my program. www.geocities.com/wavewash
Hope this helps
~Wave

Here is a nice article w/ a snippet of the AI architecture used in The Sims.

In a nutshell (quite a big one actually), the designers flipped the typical rules implementation 90 degrees and imbedded the behaviors in the world objects, rather than the characters. E.g., a refrigerator has a set of behaviors associated with it that it broadcasts when a sim is in a certain range.

This takes a bottom-up approach to ALife, something that enables designers to drop in added functionality w/o a need to redesign the whole engine.

MSN