Some thoughts on categories

I was leafing through the CMU directories on AI software and related and found a classification of AI issues:

dai/ Distributed AI
discover/ Machine Discovery and Data-Mining
doc/ Documentation and other AI-related notes.
edu/ Educational Tools
expert/ Expert Systems and Production Systems
faq/ AI FAQs: Frequently Asked Questions (FAQ)
postings for comp.ai
fuzzy/ Fuzzy Logic

It became clear to me that probably most classifications out there are going to mix *methods/techniques* (e.g. ANNs) with *natural phenomena* (e.g. reasoning) and *other issues* (e.g. "the symbolic-subsymbolic debate"). This is precisely *not* what we want the message type ontology to get tangled in. We want MeTOCA to address only *natural phenomena*, bounded by psychological and brain research.