This is something that’s said a lot, but I feel it’s become more of a simple mantra than actual words of wisdom these days.
When programming, developers may sometimes be inclined to concern themselves with optimization, even though the application is far from finished. Generally, this more strongly stems from a personal desire to optimize, rather than a need for the target platform. When people see beginners concern themselves with small amounts of optimization (micro optimizations), they tell them “stop that, no need to optimize yet!”
That’s good advice and all, but I feel it steers people away from thinking about performance at all. If you spare it absolutely no thought during development, then you may end up with an architecture that inherently performs badly. Hell, I’d say it’d be better to say “don’t optimize that, but make a note of it”. That way you can always return back to lines of code you thought you could optimize, shall it ever be needed.
Me? I’m currently keeping a watchful eye on performance. Not making any changes relating to it, but every so often I test my software on an old device to see if it still runs smoothly. If it does, then all is fine with the world. If it doesn’t, I have made not of plenty of things I can do to give the program a boost.
But first, I need to find out where the bottleneck lies. Important!