I think I’ve had to implement one complicated algorithm that was low level and not a complex business requirement in years and that was the “shunting yard” algorithm to convert a string algebraic expression to a number as part of the parser for the compiler I was maintaining.
Most developers could go there entire successful career without ever knowing how a compiler works.
Only a smattering. I could cover just about all of it in 2 nights of instruction. However, that doesn't mean the students would have mastered and internalized the concepts to the point where they'd make the right realizations when they need to.
I think I’ve had to implement one complicated algorithm that was low level and not a complex business requirement in years and that was the “shunting yard” algorithm to convert a string algebraic expression to a number as part of the parser for the compiler I was maintaining.
Right. Most of the time, one implements simple algorithms, using hashmaps and "vectors" as building blocks. But first principles knowledge can also make one better at that.
Most developers could go there entire successful career without ever knowing how a compiler works.
(Their.) Right. Most people can manage. Just like most mechanics can manage to muddle through learning as they go. But why not have mechanics learn basic knowledge about electricity? They do, because that kind of 1st principles knowledge can save people from pain. The same applies to CS/programming. Also, the knowledge itself is pretty neat. A lot of people get pleasure from learning it and being able to understand their world in a deeper and more nuanced way.
People going to top-tier 4 year universities, getting a 3.75 GPA, but coming out with no more than a glue-crud-apps-to-cookbooked-machine-learning/buzzword-of-the-day-library level of knowledge just strikes me as a colossal waste and colossal rip-off. It's also people becoming the victims of low expectations. I know that kid who tried to tell me that a pointer doesn't take up space could well have learned that stuff, if his program had just had slightly higher expectations. I'm reminded of schools who track certain kids into learning nothing more than what's necessary to balance a checkbook. That's killing young minds with the bigotry of low expectations. We're probably already at the point where some say asking kids to learn enough math to balance a checkbook is asking too much. Something about that strikes me as selling people short.
Just as one example, the basic level of algorithms, where one could realize when they're creating a O(n^2) routine, is relevant to just about every shop I interacted with working for a language/VM vendor. I found it disturbing, the number of times I could play the hero just by knowing that much.
I've worked in shops using C, where I kept bumping into workers who hadn't the foggiest idea how a C compiler might work, and it not only showed in their code, it even showed in management decisions. (Long story short: The application literally had 100's of re-implementations of linked list.)
I've interviewed recent grads from top-tier schools with near 4.0 GPAs, who try and tell me things, like adding a pointer to a struct incurs zero memory use. Then I ask them how much memory a pointer would take up, and they can't give me any kind of answer. (Or ask me a relevant question for more info, so they can answer.)
All this stuff could be avoided if there was just a certain level of basic background knowledge. They're like car mechanics who don't know the first thing about electricity. There's a certain level of background that can keep you from inconveniencing and hurting yourself. You don't need it most of the time, but when you do, it can save you a lot.