Re: CVS update: subversion/subversion/client main.c

> Ah... rereading your comment now that I've written the
> formula. You're saying that formula converges on 2N, right? Ah. So
> yes... O(N) memory. The copying is something like O(N log N)
> though. Those first ten bytes get copied a bunch of times.

The time taken is also O(n); after all, no piece of memory is written
to or copied from more than once, so the time taken can't be more than
a constant factor times the amount of memory used.

> Back to your original point: can you think of a goo term for
> describing the behavior? [that I can slap into that comment?]

You're just trying to save a constant factor, so there's no special
mathematical term. I think "suboptimal" is what you're hunting
for. :)

(Although if the file is small, maybe we shouldn't be caring.)
Received on Sat Oct 21 14:36:21 2006