Heap operations? Bipartite graph? Implementing these is vanishingly rare. I'd wager no more than 0.1% of graduates who've seen these things ever have to maintain code that implements them, never mind writing them green field. And that comes from a guy who used to maintain a compiler.
You make a convincing case for your insularity IMHO :)
Now the didactic exercise of implementing a handful of data structures and algorithms like these from scratch has a lot of value. And knowing they exist, so you can look them up and find a library / implement if necessary, is also valuable. But knowing them 10+ years after college? The details are simply irrelevant; other things are more worth knowing off the top of your head.
It's not about being able to implement them, it's about understanding the implications. Yes, I can do arithmetic by reaching for a calculator, but if I had to use a calculator for every single blesséd piece of arithmetic, I would find it impossible to do any kind of significant algebra.
Having the basics immediately to hand, without having to look them up, is needed to move on to the next level. Knowing how to recognize a bipartite graph means that when there's one around you're likely to think - hmm, I wonder if that's bipartite? recognizing the problem to be solved sometimes requires that you have a deep and fundamental understanding of the types of solutions that exist.
People seem constantly to say this - I can look it up on Google, why do I need to know this? I constantly solve problems my colleagues can't (and vice versa) because I have a deep familiarity with a range of techniques that they can look up any time they want, but don't recognize "in the wild."
There is a difference between knowing what a meal tastes like, and how to cook it.
What you are arguing for is knowledge of what meals taste like (e.g. two primary taste dimensions being space and time). I think this is really useful; so I agree with you.
But what Colin (and I mean OP) is asking for is knowledge of how to cook two specific meals, themselves rarely asked for.
I personally think CS education is mostly useless. Big-O notation and a survey of the field (a taste test of the meals, if you will) are all you really need bring along with you from discrete math, in my very humble opinion.
Cooking some of the meals while learning will teach you the general skill of following a recipe; a few more applied challenges will develop your improvisation skills. But the more applied it gets, the less CS it is.
The field is too broad, IMHO, for deeper knowledge of only a handful of things to be very useful. I know a lot about hash tables, for example, but that's because they were very useful in my job, and critical for performance. It would be a waste of time for me to know as much as I know about them now, coming out of college, never mind 10+ years later. I wouldn't come down like a ton of bricks on a recent graduate for not being able to name 3 different collision resolution approaches off the top of his head, never mind someone 10+ years out; even for a job that required their use and impromptu implementation, as mine did.
Have you considered that it's teaching you a way to think outside of your normal range, so that you can comprehend and imagine more complex designs that are easier to build and maintain?
You make a convincing case for your insularity IMHO :)
Now the didactic exercise of implementing a handful of data structures and algorithms like these from scratch has a lot of value. And knowing they exist, so you can look them up and find a library / implement if necessary, is also valuable. But knowing them 10+ years after college? The details are simply irrelevant; other things are more worth knowing off the top of your head.