Joho the Blog
An Entry from the Archives

As right as Clay is--I agree with the conclusions--he's also congratulating himself for disproving trivial examples (to borrow one of his phrases).

I barely remember syllogistic logic, but I think Clay's knowledge is fuzzier than mine. He presents "People who live in Brooklyn speak with a Brooklyn accent" as if he doesn't know what a universal quantifier is.

And he probably missed the irony of this statement: "we almost never use actual deductive logic." Uh, did he etch all 3600 words into a stone tablet, or write them on a computer and send them out on a network? Oh right, computers don't use logic...

Clay's got it right. Especially the part about making trivial things seem hard and the inability of this group to really demonstrate the utility of Semantic web applications.

I can also vouch for the "gene" example. In the lifesciences, I've watched the silly battles over the "best" representation of gene, etc. and they seem to make the smartest and most earnest people seem really out of touch compared to the people who are writing today's software. It turns out "best" is what is simple and what gets used - usually through open source - as is the case with the Gene Ontology. But biologists who never programmed cannot help but bash such ontologies because they're always missing something that they feel is a crucial to _truly_ understanding a gene, a protein, a disease, or an organism. Of course they are! It is their nature to be an incomplete abstraction, and we don't need to unify abstractions - just let people vote by using it or not (or building your own).

The Semantic Web truly is a new haven for AI types, and these folks are still challenged to show the true relevance of what they propose as Clay states.

I still think there's plenty of room for your views on this, David - I'm afraid Shirky's piece doesn't hold water very well. Jay Fienberg has a list of crit at:

http://icite.net/blog/200311/semantic_systemantics.html