"Always (notice my stipulation of 32 bit). However, regardless of the size of an integer (in bits) there is _always_ a fixed number of bits that can represent it in the manner I explained. "
Ther is no 32 bit requirement. When you are talking asymptotically, as n grows infinitely large, there is no fixed numbers of bits that can represent a number. The integers are an infinite set. > Oh, and one more thing, O(1/n) is 0, so your reductio ad absurdum > fails. What do you mean by this? I was not formally arguing the point... nor did I ever say anything about O(1/n) -- is that what you think the function is that relates the size of your storage to the size of the input? just observing that the space required for your solution grows with the input size -- thus it is not constant. What you were saying (asymptotically speaking) is that you could make a constant size hash table that could store an infinite number of values. This is obviously wrong... Your solution does not work. It is not constant space -- maybe it is constant space for a given size of integer, but there is no such stipulation in the problem. Further, if that constant changes for a different size input, then it is not constant. this is like you are saying: "3 is a constant." well, that is true, but 3 is also Log base 2 of 8. The next constant you will choose if you change the size of the inputs will be a function of the input size, as well. Therefore it is not constant. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To post to this group, send email to algogeeks@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/algogeeks -~----------~----~----~----~------~----~------~--~---
