XP Home

Optimize Last Because it May
Not be as Slow as You Think

Lessons Learned

 We needed to add some new functionality to the VCAPS project at Ford. Based on some other code in the system that ran 8 hours this new section would run in no less than 16 hours even after we squeezed ever little bit of performance out of it. We added the new functionality as simply as possible. We ignored concerns of speed in favor of concerns of code clarity and maintainability. Make it run, make it right, make it fast.
 When we had finished the code we fired it up on a Friday afternoon in anticipation of several days of run time since we had not done any
optimizing. We were shocked when it came back in 1 hour. By keeping our design simple and understandable we had managed to avoid several "penny wise, pound foolish" types of optimization that other sections of code had fallen prey to.
 At this point we could have considered optimization, but because it was so much faster than any other portion of the system we didn't need to do any optimization at all.

Don Wells
eXPerience Software

ExtremeProgramming.org home | XP Lessons Learned | Optimize Last | Email the webmaster

Copyright 1999 J. Donovan Wells all rights reserved.