Re: A new proof of the superiority of set oriented approaches: numerical/time serie linear interpolation

Cimode is arguing that as long as it works, one query is as good as another. If we were in a procedural language, Cimode would be saying "Bubble sort, Quicksort! what's the difference?"

Selzer is arguing that the complexity of one query can be higher than another -- algorithmic complexity. If we were in a procedural language, Selzer would be saying "Bubble sort is O(n^2) and Quicksort is O(n ln2 (n)), that's the difference!"

Selzer is right. Complexity happens to effect execution times, but it is more abstract than just reading a stop watch. At some point, you cannot throw hardware at a problem and need to use another tool.

>> .., so I had to improvise in order to implement a transition constraint.<<

I have an article in www. dbazine.com on implementing transition constraints with auxiliary tables and DRI actions. No need for triggers. Received on Thu May 03 2007 - 14:58:47 CEST