If there’s some particular code that they read once and found awesome that says a lot right there. They’re passionate about writing code. If you take their enthusiasm about said code at face value, that is.

I’m not sure how to feel about the count the bits sort of question. Recent grads seem to do consistently terrible with questions like that. My gut tells me that it’s because they’ve gotten incomplete educations, or are just not particularly strong as individuals, but I’m not so sure. When I went to school, bits and bytes were pretty fundamental and were well covered. How can you write tight code, if you didn’t really understand what was happening “on the metal” … right?

In the early 90s, I wrote C and ASM code that used this knowledge, called APIs that used flags encoded as bits, etc.

But I haven’t seen code like that for a long while. Talking to experienced, active developers, they tell me they haven’t either. Eschewing highly efficient storage techniques, such as using bitfields because RAM is cheap seems both a bit wrong but consistent with trends like Vista needing 2G RAM to run well. Higher level constructs may make for easier to read/maintain code, I guess.

So is it not important to teach? Are languages and APIs high-level enough that you don’t need to know it - and therefore it’s a waste to teach in a finite curriculum? I’m not sure.

I am sure that I’d be more comfortable with someone who was so “into” programming that they learned that kind of stuff on their own.

I agree with everything you said, though, and it’s cool to see your techniques evolve.