Another concrete example of the conflict is classical vs quantum physics which is nearly exclusively taught as the historical story of the quantum revolution or whatever you call it around 1900.

It would be trivial to skip the historical story and just "yo here's a FET transistor, a solar cell, and a vacuum tube, and the math just works, so get used to it" but nobody is ever taught that way.

Would be interesting for someone to attempt to write a non-historical physics textbook.

On the other hand, in IT, this style of learning is considered normal. "This is how the kool kids do things today. The past does not matter and probably didn't exist anyway. The end."

> Would be interesting for someone to attempt to write a non-historical physics textbook.

I recall some prominent physicist saying they've never read Einstein's original papers on relativity; and they don't have any real desire to, since we understand it much better now, and have better examples, notations, etc.

Most physicists have never read them, nor would it help. Most classic papers go unread, and it’s not a bad thing. How things get sorted out is usually a huge mess and following the real history is an enormous distraction (and can confuse and misdirect students).

While that is true, I would highly encourage everyone to read the original EPR (Einstein Podolski Rosen) it is so amazingly clear in its writing and can be understood even by laymen.

A downside to the modern approach is that “plug the equations into a computer and let it churn for a bit” plays a big role I think. This is seems a bit annoying from a teaching point of view

* “here’s what the computer came up with” doesn’t really build as much intuition as doing the equations out by hand

* there’s enough material in “how to plug it into a computer and make it run quickly” for a whole degree

>On the other hand, in IT, this style of learning is considered normal. "This is how the kool kids do things today. The past does not matter and probably didn't exist anyway. The end."

And yet nand2tetris is one of the most common suggestions for people who actually want to figure out how computers work. Because you actually are posed with problems that real world engineers used to struggle with, and are allowed to play with the problem before finding out the answer. It builds a much deeper understanding than "it works like this because it is"

But nand2tetris doesn't have a historical approach if i'm not mistaken.

“How do we know the world is round” is an interesting example to use. Eratosthenes got an estimate for the radius of the Earth that was pretty close using, I dunno, sticks and his eyeball or whatever ancient Greeks had.

The history is useful there, I think. First off, it is motivating — that’s a pretty big question, really shows what math can accomplish without much hardware. Second, it provides some sort of useful constraints if you are trying to work out the computation yourself. In a textbook the problems are always a bit artificial, some info will be given, some hidden. “Anything an Ancient Greek would have” is a bit more real-world, you can see where the constraints come from.

Erastosthenes did it by comparing the angle of the shadow cast by a stick on the summer solstice at noon in two cities, Alexandria and Syene (now Aswan), which are about 800 km apart. What was special about Syene was that it is so close to the Tropic of Cancer line that the sun would hit the bottoms of wells on the solstice, meaning that he only needed one stick, in Alexandria, and the distance between the cities, to do his estimate. His error was about 2%.

He's also famous for the sieve of Erastosthenes for finding prime numbers and figuring out the earth's axial tilt.

Darn, at some point I started imagining he had a protocol for synchronizing clocks over a distance. assuming the difference was in longitude. (E.g. Symmetric NTP over carrier pidgeon.)

Longitude was a much tougher problem to solve.

Yea, recalling the book of that name might have contributed to my drift of recollection. Solvable earlier on land (but not sea) on terms of clock synchronization.

You can get a vaguely credible estimate by standing on a (small!) hill and measuring the distance to the horizon. Works better if you have a body of water handy, because bodies of water are nice and flat.

I think it can depend on the particular discovery whether we should teach "here's how the first person figured it out" or skip right to "here's how a later person found a much better, easier to understand approach".

I've been interested for awhile in a distinction between math and philosophy here. Just descriptively speaking, we don't tend to teach math from original source (at least, not well-established math like you'd see in high school and early college). We don't teach calculus out of Newton/Leibniz/Cauchy/etc, we don't teach Fourier analysis out of Fourier, etc. Maybe geometry students will work from the Elements, but I certainly didn't.

But in philosophy, students do read everyone from Plato to Rawls in the original. There are (lots of) supplementary texts, of course, but they're companions.

I think that shows an interesting divide that the way we actually think about philosophy is not just a catalogue of arguments and positions one could take, but that the famous works are things worth reading unto themselves. There are a few math papers like that ("God Created the Integers" tries to anthologize them) but they are the exception rather than the rule.

But maybe there's another explanation?

Two contributers I can think of

1. Being able to tease apart language from ideas can't be 'taught' you have to do it

2. People disagree much more in philosophy than math