...well to start right in on a probability theme, the question came up in discussion in the lobby at SFI's annual meeting about Fuzzy logic.; Specifically its use in modeling. The ML community has largely discounted its use, and the current texts just mention it in passing.

The point is to find a balanced comparison of Fuzzy and probabilistic methods. I know what you're thinking already -- so stop before you say it! I'm not advocating anything. Its just merits a careful comparison of what is lost by Fuzzy approximations and what the gain is, in simplicity.

I'll have to go back to ``pre-wikipedia'' sources (e.g. books) to work this one out.