Perhaps the more important practice is to establish visibility into which areas of your code are the performance bottlenecks. If you're developing in an environment where you're comfortable quickly profiling things with low friction, then it's much easier to know where to spend your time optimizing data structures and algorithms.
This is now one of my favorite technical talks. Very salient points!
I absolutely love his section on how calling some programmers lazy when all visible evidence shows otherwise is silly.
Re:Choosing the right data structures
by candydance on Wednesday October 23, 2013 @10:12AM
(#6598819)
This is a pretty awesome talk. I wish he expanded on the "array of records" thing and how he implemented it.
Maybe an array of key-value pairs?
> Using the right data structure is usually bad
He argues, because really, most of the time it won't matter.