lua-users home
lua-l archive

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index]


On Tue, Mar 22, 2011 at 11:28 AM, liam mail <liam.list@googlemail.com> wrote:
> In my opinion if speed is consideration for a project, you can reproduce the
> same results as the page shows and you heavily use OO then pick the one
> which matches your expectations.  In these situations speed should be a
> consideration from the start and not an after thought.

Not quite, if speed is an issue you consider from start for
algorithmic efficience. Like log(n) instead n, etc. But all fine
tuning should be done after profiling. I can tell from several of my
projects, the actual profile came to a hugh surprise to me, instead of
the places I tried to optimize when coding. As the famous example
code, if the majority of your code only uses 5% of the total runtime,
even if you super-optimize it to zero, the total impact will be
minimal. More often than not, for me the actual profile turns out some
unimportant sounding routine taking >50% of time. And if you
concentrate all your energy on that routine, the impact will be heavy.