3 Jun
2009
3 Jun
'09
2:29 p.m.
I'm questioning the degree of assurance required for a new algorithm to be unleashed on the unsuspecting masses.
Exactly the point I was trying to make. To be more precise the novelty of an algorithm shouldn't be held against it. What matters is that a reasonable degree of assurance regarding correctness and performance can be given. This concerns the library as a whole, not just whatever algorithm(s) it may use. I understand the concern of not wasting time evaluating libraries using algorithm(s) that are incorrect in the first place. -- EA