[
Lists Home |
Date Index |
Thread Index
]
On Thursday 10 January 2002 08:23 am, Al Snell wrote:
> Although that kind of stuff is often quoted as being rampant, I
> instead am always finding that it's hard to optimise things late in
> the game because people have often come up with designs that make
> certain classes of useful optimisation hard...
...or come up with naive designs that simply don't work on a larger
scale.
This is the art of programming, where experience, and common sense can
make the difference between having constant rapid evolution vs.
controlled evolution.
Both modes are useful, but controlled evolution often costs less in
the long run... and in enterprise, the difference can add up to
millions.
(... for those XP'ers out there, no flames please. People that know me
know that I'm not a "design the perfect thing up-front" person...)
|
