Subject: Re: [boost] Overload resolution speed
From: Mathias Gaunard (mathias.gaunard_at_[hidden])
Date: 2011-09-23 08:16:40
On 22/09/2011 23:40, David A. Greene wrote:
> Dave Abrahams<dave_at_[hidden]> writes:
>>> In general, all algorithms in a compiler should be linear, or worst
>>> case n*log(n). Any quadratic or worse algorithm is pretty much a bug.
>> I'd like to think so, too, but I'm not sure all implementors would agree
>> with you.
> The problem complexities are not linear or n*log(n) so why should you
> expect the algorithms to be?
I'm pretty sure Sebastian was talking about front-ends.
Nothing in the front-end should be worse than n*log(n).
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk