Hi Darin.

19-May-04 17:29:58, Darin McBride wrote to Pascal Schmidt


 DM> Hello Pascal!

 DM> Replying to a message of Pascal Schmidt to Neil Heller:

 PS>> What I'm saying is that sometimes an algorithm that looks faster
 PS>> on paper turns out to suck in practise.

 PS>>>> The most primitive algorithm for this problem (compare each
 PS>>>> with each) would be O(n^2).
 NH>>> Yipes, I hope it doesn't come to that.
 PS>> That one is very memory efficient, though, and good enough for
 PS>> smaller input sets. Think about using your approach for 64 bit
 PS>> integers, how much memory would your hash table consume?

 DM> With a reasonable hash algorithm, not much more space than the
 DM> original list ;-)

Which is fine as long as the ordering of the output is non-critical
(hmm I guess you could use an O(n) sort on the output. :)

 -=> Bye <=-

---
(3:640/1042)