On Thu, Jul 19, 2012 at 01:15:15PM -0300, Zooko Wilcox-O'Hearn wrote: > On Thu, Jul 19, 2012 at 1:14 PM, Zooko Wilcox-O'Hearn <zo...@zooko.com> wrote: > > "There are only two hard problems in computer science: cache > > invalidation and naming things." -- Phil Karlton 1947—1997 > > P.S. The fact that it is the *invalidation* that is the hard part is > probably salient to your proposal here. :-)
Well, as shown by the PGA for OLAP materialized view selection (and the ensuing 15 years of research papers about how to improve it), sometimes even figuring out what to cache in the first place is another hard part. My proposal conjectures that a global solver will be able to do a substantially better job of that than the existing solutions, which often harm global performance in order to optimize the performance of one component. Kragen -- To unsubscribe: http://lists.canonical.org/mailman/listinfo/kragen-discuss
