granting funds based on pleasing?

[rjf]

Just thinking, if one were to establish a foundation to encourage

the study of pure mathematics based on the principle that

such study has the sole effect of bringing pleasure to mathematicians,

how would one rank proposals?

Here's my (initial) thinking:

A (1 to 10): how much pleasure per mathematician in the relevant interest group

          (on average) would this study provide? [1= not much. 10= lots]

B (1 to 10): how many mathematicians are in the interest group? (maybe log_e(n+1)?)

C (1 to 10): how important is it to give pleasure to this group (e.g., are they

 famous?) (totally subjective?)

Then different proposals' rankings would be computed based

on the product A*B*C.

Compare to the NSF's traditional ranking which is supposed to reflect 

intellectual merit

  as well as

broader impact.

In my experience it mostly reflects the prejudices and views of

the handful of reviewers (+ NSF program directors) who will

choose the winners by whatever process they have internally,

and then re-express them in terms of merit/impact as necessary

to get through the process.

Now it is possible to perturb this mechanism by awarding monetary

prizes  (increases pleasure usually).  E.g. 

Clay Institute prize of $1 million for Birch and Swinnerton-Dyer Conjecture.

It can go both ways though. If you spent your life studying this

and someone else got the prize, that might negatively impact

your pleasure.

I have only once sat on a review panel in the Math directorate,

so I can''t say much from experience. However,

  I'm not at all convinced that A*B*C  isn't the way it works

behind the scenes.

RJF