Hacker News new | comments | show | ask | jobs | submit login

Great topic! It reminds me of the times when I found out about the wonders of profiling my applications, where I could clearly see where the bottlenecks were (the 3%). This also reminds me about the endless discussions about which <insert here a programming language>'s function/method was faster than the other and why you must change your code immediately to save a tiny fraction of a millisecond </irony>

I haven't read the book yet, but I believe the key to a mature optimization is knowing clearly what needs to be optimized. There is no way to know it if you don't measure it. But the thing is that you may be measuring the wrong thing. If you focus on a class, for instance, without the entire context, you may be trying to speed up the wrong piece of your application. Make sure you measure the entire flow, from the user clicking something to him seeing the response. Sometimes you will find that what needs to be optimized is not even your code.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact