On Thu, Sep 6, 2012 at 11:37 PM, Johannes Bauer <dfnsonfsdu...@gmx.de> wrote: > Not in my original post. If you read it again, you will clearly see that > I was talking about purely random strings. And since you like to > nitpick, I'll clarify further: I'm talking about bitstrings in which > every bit of every character has the same probability of occurence, 50%.
That sort of string isn't a normal thing to be comparing, though. Here's an idea. Someone who's doing a lot of arguing in this thread should take Python, find the string comparison routine, and hack in some statistics-gathering. Then run *real code* on it. Maybe use this with one of those web frameworks and run your web site on it for an hour or two, or fire off some real scripts you really use. Then dump out the stats at the end. My guess: The bulk of string comparisons that get to actually comparing byte-for-byte will end up returning True. Most of the false comparisons will be proven earlier; if I understand correctly, Python will check for identity (easy true) and different lengths (easy false). But my guess could turn out to be flat wrong. In any case, it'll be far FAR more useful than arguing from totally random, or random word selection, or anything. Who's game? ChrisA -- http://mail.python.org/mailman/listinfo/python-list
