So .. I read it up to the point I think I get what's going on ..
Things I like:
- use of cellular automata style of algorithm/programming to do AI or goal-seeking in games; looks like they've found an interesting approach where you get what you want entirely through emergent behaviour
- AgentSheets (
https://en.wikipedia.org/wiki/AgentSheets) looks like an interesting product, and I'm confused how I've never heard of it; it has
if(event) then(action)
style rules for agents operating concurrently in a grid and can do 2D and 3D games this way (
https://www.agentsheets.com/img/game3.png▾
)
Things that leave me with uncomfortable feelings:
- AgentSheets is really, really old (see the website, all its links are broken and it looks pretty bad -
https://www.agentsheets.com/store)
- it looks to me like they're trying to re-invigorate their product by coming up with this radical-sounding concept of "anti-objects", but there's nothing there: just because you use parallel agents to implement a goal-seeking algorithm doesn't mean you've discovered some amazing inverted programming paradigm!
By the way, the author of this paper, Alexander Repenning, has managed to get millions of dollars in funding for his (related) work from the NSF (
https://www.nsf.gov/awardsearch/showAward?AWD_ID=0205625 and
https://www.nsf.gov/awardsearch/showAward?AWD_ID=0349663) which should inspire folk on this Slack!