I care a great deal about the performance of my programs, but I try to spend my time wisely. There is a huge difference between studying and modifying a working program in order to speed it up, and worrying about very small things before your program is working. I do the former frequently and encourage it, but try to avoid the latter. The most obvious reason is that it's very hard to predict where your bottlenecks will be before you have a working program, and you may end up making a section of code very hard to read (and spending a lot of time on it) for little or no gain. So, optimize away, but don't do it where it isn't needed.