On Wed, Jul 04, 2007 at 10:56:57AM -0400, Christopher Browne wrote:
> That doesn't change anything about "agreeable ordering" as far as I
> can see.
> 
> Of course, that example isn't much of a "win."  What would be way more
> interesting (from a performance perspective) is the case where there
> are 25 deletes in a row from my_table that could be folded together.

Right.  But the actual key here is to make sure that (1) no agreeable
ordering actually changes _and_ (2) that doing this work for every
sequence is in fact going to be a winner for most cases.  Jan and I
talked, IIRC, about this sort of optimisation in the early days of
1.0 design work, and neither of us were able to come up with a set of
tests that could both demonstrate this was a win _and_ that it
wouldn't kill ordinary-case performance (or blow out memory).  I'm
not saying it's impossible, but I'm trying to suggest that some
general algorithm work in this area is going to be needed, and it's
not going to be ameable merely to empircal tests.  You also have to
have at least _prima facie_ evidence that, for ordinary cases, this
won't suck.  I'd like to see that argument.

A

-- 
Andrew Sullivan  | ajs at crankycanuck.ca
When my information changes, I alter my conclusions.  What do you do sir?
		--attr. John Maynard Keynes