>I'm saying, programming often is just knowing which functions to pick, which is not an inert skill except for reading comprehension, if there is a good documentation. There's nothing universal to learn there, the interesting bit is writing those functions. But a pseudo random number generator is rather advanced stuff.
First of all, I'm not asking them to know "advanced" random number generation theory. As a kid, using random number generators in BASIC was one of the first things we did when we learned programming - it makes writing programs a lot more fun. Granted, it's not usually covered in introductory programming courses. But all that is needed is one API call. I pretty much told him what he needed to do. I showed him the API he needs to generate random numbers. We talked in big picture terms about the algorithm (set up the experiment, run 10000 times, take the ratio of success over number of runs). We talked about the theory on why such an experiment would answer his question. I helped in every way other than the actual coding. Had he written some code, I would have helped debug (some probability experiments are tricky to code correctly - that is acceptable).
He just couldn't translate the problem to code.
>You are being very critical, a part of university education is tolerance, although some opinions differ on if that relates to knowing literally everything. Trade-offs are a part of engineering because optimal solutions don't always exist.
I definitely am being critical. When you graduate, you get a seal from your institution that is supposed to be a guarantee of something. My question is: What is that something? If you start dealing with math graduates from a university and you find they often cannot do basic algebra, you stop valuing the seal the university provides - which damages all math graduates of the university. They are working very hard to get a certificate that is not valued by others.
Would you trust a physics graduate who cannot do basic calculus? You can argue that he may know all the theories (conservation of energy, electromagnetics, etc), and understands the principles behind calculus (area of curves, rates of change, etc). But if they struggle with basic evaluation of integrals, you probably will wonder about what kinds of courses they taught in his university that allowed him to get a degree without actually solving problems with calculus.
This is the problem the author is talking about. If you consistently interview people with computer science degrees, who cannot do a simple fizzbuzz, you stop valuing the degree - and that is the general landscape in the software world. Those who worked hard to gain a lot of knowledge relevant to both theory and practice are right to be upset that because of lax (or at least misguided) standards from their educational institution, they have to go through more hoops to convince people of their skills.
It goes beyond the simple mantra of "Universities teach computer science, not software." I would question the technical ability of anyone who cannot write such a simple program, with as much guidance as I provided - be they a CS major or a math major. I'm not asking for practical knowledge like whether unit testing is useful, or how to architect large software, or the virtues of OO over FP. Just a basic program. Can you write it?
Certainly there is not universal agreement on the curriculum. I'm saying there should be agreement on some basics. Historically, universities handle this by ensuring some introductory course is in the curriculum. My argument is that in the US the education system is a little too modularized, and there should be more coupling. If programming is seen to be a core skill that engineers should know, make writing code an aspect of every engineering course (it's really easy to construct assignments/projects in pretty much any engineering course that could involve writing code).
Don't let someone get an A in the introductory course in his freshman year, and then allow him to forget pretty much everything he learned from it by the time he graduates. If you do, you wasted the student's time, the instructor's time, and the time of all the interviewers who cannot trust your degree and have to check for themselves if this person knows the basics.
Again, it all boils down to: The university is giving you a degree. Is this degree a guarantee of anything? If so, what? I'm all for education for the sake of gaining knowledge. But how many of us would be OK with gaining the knowledge without any kind of certification from the university? Not many, I suspect. I actually did this personally (dropped out of PhD after many years as I felt I gained the knowledge but did not feel a degree would help me), and can assure you from the conversations I had with others who tried to convince me otherwise that pretty much almost everyone wants that paper to have some value.