It’s an interesting quote, however, I think people misuse it. I always review the code that I write and optimize where I can without complicating things. Good design is important, and if you’re making major changes to your design when you optimize you obviously did not spend enough time planning out your design.

That is, however, not what I did when I optimized the code. I looked over things and noticed a few places where I was needlessly recalculating things. So I removed them or moved them out of for loops. This significantly sped up the generation time for certain fonts.

Personally, I believe any good software person should review their code after they write it. I don’t think anyone can write perfect code all the time on their first try. And if they claim to do that, they’re just full of BS. Reviewing your code and fixing up certain areas is a good policy. It’ll result in better performance and a better end result for the program user.

Interesting article on the subject:
http://www.acm.org/ubiquity/views/v7i24_fallacy.html

A book on PHP really isn’t necessary unless you’re going to be doing some heavy duty stuff or some database work. I actually don’t own one (though I may look into getting one). I just took a few tutorials online and use the online documentation to look up the information I need. I know what you mean about not wanting to get into another language though. I think sometimes I try to do too much stuff at once and then end up not doing much at all (example: last Fall I picked a book on LISP and Java EE, I’ve only read a little from each one).