[ros-kernel] new patch

[Steven Edwards]

Hiya Jason,
--- Jason Filby <jasonfilby at yahoo.com> wrote:
> Just a note here: sometimes, well quite often really, its better to
> do something right the first time. Architectural decisions are
> especially important to be done right first, but other code too. This
> is because the issues at hand are fresh in everyone's mind and that
> old code usually doesn't get fixed up as much as new code gets
> written. Also, new code gets written around the code thats being
> discussed now - so that its sometimes is quite difficult to fix
> without having to rewrite the new code too.

Right I agree with that.

> I'm not saying that this code shouldn't be committed; I'm talking of
> coding in general.

Right. This is a hard case as the code does fix the problems pointed
out. Its mainly a question of design and performance. The standard we
have followed to date is get it to work first and optimize later. I
think this is still a good method to use but maybe not always. Mozilla
is a good example of this. They have a source file that is over 10,000
lines long because they didnt have a good design plan when they
started. 

Still in this case I suggest commenting the hell out of the code so at
point point down the road we can look at speeding things up. 

Thanks
Steven