>One of the problems we developers face, is that many of us have the latest and greates Dual Core or Quad Core machines with loads of memory. We develop on these machines, and it works pretty fast and seems responsive. However, once we deploy it out to the field, we start getting complaints about performance. So what do we tend to do….test performance on our nice fast snappy development machines. Unfortunately this does not necessarily produce the desire result, because we are not testing against what our user base is using.
Performance is a hard thing to measure. It can vary by operating system, environment, and computer hardware. A small change in the way you write your code, or the amount of code that has to be executed can make a big difference in a different environement and operating system. So while there may seem to be no difference in execution time on your high powered development machine, that one less statement that is executed a 1000 times can make a world of difference on your user’s machines. Fewer uses of instanceOf, or Casting can help as well. These seem relatively harmless but their costs can show up in unexpected places.