As I'm building applications I find myself constantly asking if this is best way to perform or implement a certain functionality. Often, I'll post questions on stackoverflow or another forum desiring feedback only to receive comments about how not to "put the cart before the horse" regarding performance. Do most programmers really not think about performance until the application is finished, or performance is absolutely unacceptable?? I mean, I understand that development environments differ from production environments and that you shouldn't completely rely on the results from your dev laptop...but, there are practices and techniques that yield better performance than others.
Is it bad practice to consider performance throughout the development process? Should I push these considerations off until performance actually is tanking??
Update
Just to be clear, I'm talking about the situation where you are considering or just about to work on some piece of functionality. You know there are several ways to implement it, but you're not quite sure how well each implementation will scale. Also there might be several techniques you're not even familiar with. On a small scale any of the approaches would probably be adequate, but on a larger scale some will keep up and some won't. Often when I ask for opinions or guidance the response is: worry about it later...