logoalt Hacker News

driveby_cat_tal • today at 4:42 PM • 0 replies • view on HN

I agree a lot with your take about people not wanting to put effort into understanding a complex concept and blaming the definition, but I do think your specific example of compactness and the definition you focused on potentially illustrates the other side. For me, I think the point of compactness is easier to understand in terms of sequences. A big part of why compactness as a definition matters is because it specifies when a sequence will "run out of room" and be forced to converge to a point, so you can take limits and do calculus things with it. In finite dimensions, closed bounded sets are always compact, so it's hard to intuit why this matters. When you get to infinite dimensions, it's no longer true. You don't run out of room because you always have a new dimension you haven't been to to take the sequence in. That's where you see more of the value of a definition like "compactness" - it distinguishes where closed and bounded sets let you assume you're going to be able to force sequences to eventually converge and prove things with that assumption. For me, personally, encountering that sort of example helped me understand compactness much better and why I was studying it, even though I had spent a lot of time learning to prove things with the more abstract definitions. For me, it was a forgettable definition without finding the motivation for it.

I think there's a tension in how people mean "understanding". On the one hand, the example I gave (in my opinion, at least) helps understand the motivation of compactness and why anyone cared to define it in the first place. It makes it easier to understand why the abstract definition is stated the way it is, and provides some intuition that can stimulate approaches to proving things with it. Some of the sibling comments have discussed the tendency in math to distill education to focus on the top-level, abstracted definitions and discard the motivations that led to them. I think finding a way to incorporate those motivations is something that is often overlooked, at least in my experience, and may be some of what this point about "explaining in simple terms" (at least for me, in the example above it's "simpler" than the one you gave).

That is a particular form of "understanding", though, and it doesn't take away from your point. You can't take the intuitive, motivating example I gave and use that to make arguments about compactness without doing the effort of making connections to a more formal definition like the one you gave. In that sense, even if the example I gave were to help someone "understand" in the sense of getting a reason behind the motivation for compactness and what it's used for, it doesn't help them "understand" in the sense of giving them facility in actually working with it. That's where I see a lot of value in your point. Like you said, even if there are examples that may be more intuitive, there is a tendency to not acknowledge that really understanding some complicated things just takes some work.

I feel similarly about complaints I sometimes see about variable naming in programming. Descriptive names make a huge difference in some cases, but the reason a person who hasn't spent time to study differential equations and has trouble seeing the point of

dx_n = dx_p + dt * (-kx_p -cdx_p) / m

doesn't fully understand what's going on if it's written

derivative_of_position_at_next_time = derivative_of_position + change_in_time * (-stiffness_of_spring * position_at_previous_time - damping_constant * derivative_of_position_at_previous_time) / (mass)

...even if those words might help them some picture. I feel like a rough analogy for the concept might be, a person who really "understands" the concept, could probably come up with these variable names to explain the convention, as opposed to someone who has memorized the formula - but (oversimplifying at the risk being unfair to these arguments) there are also people who suggest the only reason they don't understand what the code is doing is because people don't write variables like in the second case.