Re: Bunch o' dead code in GEQO

Neil Conway <neilc(at)samurai(dot)com> writes:
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
>> I'm assuming that the original author of the GEQO code already did
>> that testing ...

> Removing the code without bothering to verify this assumption is a
> little unwise, IMHO:

Fair enough. I did a little bit of poking around and it seems that
ERX (edge-recombination crossover) is still considered one of the best
available choices for solving Traveling Salesman problems via genetic
optimization. Now there is the little issue that our problem isn't
really TSP and might behave a bit differently, but I see no evidence
to suggest that one of the other recombinator methods would do better.
They were all designed to do TSP.

The big problem here is that I don't see any practical method of
obtaining indisputable proof that one method is better than another.
Where are we going to find a representative test set of dozen-or-more-
way SQL join queries?

regards, tom lane