Re: [boost] Re: dynamic_bitset: performances and other notes - dyn_bs_modifications #1.zip (0/1)

From: Rene Rivera (grafik666_at_[hidden])
Date: 2002-07-31 20:26:50

[2002-07-31] Jeremy Siek wrote:

>On Wed, 31 Jul 2002, Gennaro Prota wrote:
>gennar> On Wed, 31 Jul 2002 16:06:33 -0500 (EST), Jeremy Siek
>gennar>
>gennar> Yes. Of course we have to decide a strategy. My idea was to make
the
>gennar> customization absolutely transparent to those who don't use it, and
>gennar> see if the code complexity could be reasonably limited. Do you
>gennar> dissuade me from any attempt, then? :-)
>
>Yes. Sorry to ruin your fun, but someone has to represent the poor guy
>who, 5 years down the road, has to fix a bug in dynamic_bitset. It will be
>hard enough to understand the optimized version, never mind the added
>clutter and dispatching needed for customization hooks.

By the way I do tend to agree with the, don't make it complex for the sake
of optimizations. But at the same time we have things like policy pointers
which are complex, it seems to be the trend.

I think it is important to try for some flexibility... after all when
someone with a CPU with parallel/vector operations wants the extra speed
they can at least have chance of having it ;-)

-- grafik - Don't Assume Anything
-- rrivera_at_[hidden] - grafik_at_[hidden]
-- 102708583_at_icq - Grafik666_at_AIM - Grafik_at_[hidden]