it seems to me that people with 10+ years of experience have more empirical evidence what is actually needed/useful.
Your comment and other similar ones sound to me as textbook material which is abstract and correct in principle but not exactly how it is when you spend some time in industry.
I have my own empirical evidence as an user: a lot of software I use is slow or uses too much memory, or has memory leaks. As a programmer, I keep wondering if those could have been averted if the programmers working on that code would have used better/more efficient algorithms, instead of just going with the simplest solution.
No. It's not. It simply inexperienced programmers. I bet they even have CS or EE degrees.
One of the biggest culprits of truely awful programs are graphics card makers. I bet they all know algos, the problem is that they think they know how to program when they're little more than amateurs or hobbyists.
