[_mocha]

I'm somewhat surprised to see this on the front page of HN. I distinctly remember this material being covered in a single lecture during our freshman CS course 20 years ago. I doubt this would've reached the front page in HN's early days, and it makes me wonder if we've moved away from teaching these fundamentals due to the heavy reliance on abstraction and ai agents in today's industry.

[asa400]

I've never taken a lick of CS in a formal setting, and I feel like that's increasingly common as programming has broadened its employment base. Most of the people I work with haven't done formal CS, either. Consequently I've had to learn all of this stuff on my own. There are communities out there that value education this kind of education. I can vouch for the Rust community, which has helped me learn _a ton_ about this kind of "lower level" stuff over the last 5 years despite starting my career doing Ruby.

[jerf]

One of the neat things about programming is that you can develop a high degree of skill in it without any formal education at all.

But it does mean that we've always had those programmers who have had characteristic holes in their education as a result.

Personally I enjoy having a mix of self-taught and formally-taught around; I think there quite a few things the self-taught are often better at than the formally educated as a group. I respect both sides. I'm just saying that there has always been a market for this sort of thing on HN because the self-taught have always been around.

[nomel]

> I think there quite a few things the self-taught are often better at than the formally educated as a group.

Would you mind expanding on this a bit?

[jerf]

First, there are people who manage the best of both worlds, so bear in mind I'm talking about the ones who only manage the one side or the other.

The self-taught ones are often much better at diagnosis of problems in my experience. "Hey, the servers are blowing up, own the process from diagnosis to deploying the fix" is something I can generally throw at a self-taught dev faster than a school-taught dev. Often because even if they've never owned a server they've already been through the "my graphics driver is crashing my desktop when I plug in my USB drive" dance on Linux a number of times and know how do go through diagnosis processes.

They're also often better out-of-the-gate at being handed a large task and knowing how to break it down quickly without much direction. Sometimes they should be getting more direction than they do; this is often the source of bits of the code base that seem to be completely foreign in style, library usage, etc. Gotta watch the PRs.

On the other hand, the self-taught will sometimes blunder into some critical O() mistakes because they haven't been taught to think that way, and may never have worked at any real scale before. They can also end up doing too much without using a library because they're used to just doing it themselves. (Not that school teaches you to "use libraries" exactly, but the school-taught are just generally less fluent and as such more likely to look for easier solutions... which is often exactly what we want!)

Where the holes really develop is when we get into heavy data structures, or a compiler-like code base (though sadly many school-taught people escape without compilers), or heavily recursive code, which doesn't just include "recursion" per se but also things like working with DOMs that are recursive tree structures. A lot of time the self-taught are great at straight-line "do this, then do that, then do that" code bases but struggle to build longer-term maintainable code that has good reuse.

And I'll say it again, these are all absolutely, positively, generalities and they do not apply to everyone. Some self-taught people are great at creating high quality, well-architected code bases. Some people escape school not just without compilers, but without any discernible skills whatsoever. (Those are some sad interviews.) These are just broad patterns I've observed.

[bigstrat2003]

I have a degree in CS from around the same time frame, and we never covered topics like this. So I think it varies widely, and is not necessarily a modern thing for people to not learn about struct sizing.