[Corpora-List] "Cargo cult" NLP?

John F Sowa sowa at bestweb.net
Fri Apr 11 16:24:09 CEST 2014


On 4/10/2014 9:30 AM, Noah A Smith wrote:
> indeed much research is (and should be) done without worrying about
> eventual usefulness in practice. Don't like my research result?
> Okay, don't use it.

Yes and no.

The best examples are publications in the 1930s on automata theory and recursive functions by Turing, Church, et al.

But there have also been cases where professors developed a "cottage industry" of getting their graduate students to prove lots of theorems whose conditions were seldom, if ever met in any actual application.

I recommend the following article by Joseph Goguen:

http://cseweb.ucsd.edu/~goguen/pps/tcs97.pdf

Tossing algebraic flowers down the great divide

John



More information about the Corpora mailing list