Wednesday, August 22, 2007

Alan Turing and the nature of the beast

Between Scott's my-paper-was-rejected-from-FOCS post and the hot-topics discussion on Michael's blog, I have seen one too many references to Alan Turing as the great pioneer, founder of computer science, who would not have his paper accepted today in our sorry excuse for an scientific community. To best describe my opinion about this, I will borrow the following image from Scott's job-talk at MIT:
To see the problem with the argument, we need to look no further than the title:

  • Alan M. Turing, ``On Computable Numbers, With an Application to the Entscheidungsproblem,'' Proc. London Math. Soc., 2(42) (1936), 230-265.
In other words, Turing's trying to solve a very explicit open problem, posed in [Hilbert and Ackermann, 1928]. His writing follows the "standards" of STOC/FOCS papers today, including a serious attempt to convince the reader that the proof technique is only superficially similar to Gödel's, and a reminder that his result is technically stronger. I see no grand claims of the sort "we initiate the study of [buzz] of computability, as a way to understand the fundamental nature of the universe / human reasoning / ... [buzzz] and creating a field that will generate countless open problems [buzzzz] ".

Let's get over the dichotomy between the misunderstood genius who spawns new fields with a stroke of the pen, and the number cruncher who generates pages of unreable technicalities to improve some bound. (The rest of the world got over romanticism about a century ago.) We're not disparaging Turing by saying he was only after an open problem!

The progress that really matters (be it on open-ended or very well defined problems) is based on an original view of the problem, plus some work that anybody could have done (given time). Any serious attempt to solve a hard problem has the "what the heck do we actually mean to prove?" component. If you want to prove you have the skill to answer such questions, there's hardly any need to invent a new field.

Our beasts (open problems) have but one nature: an appetite for brilliance.

No comments: