Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

So you want to learn physics (2021)

susanrigetti.com

Post reply on HN
Print view View original

61–70 of 170 posts

Re: So you want to learn physics (2021)

#61

Post by ecshafer »

ChuckMcM wrote: post #35
qsort wrote: post #3

A point that's rightfully emphasized by the author: > Solving problems is the only way to understand physics. There's no way around it. This generalizes well to other fields. I don't want to discourage anybody from trying to educate themselves in a difficult field (be it physics or something else), but that's a very common and immediately visible problem with autodidacts. If you haven't worked through enough hard pro…

Totally 100% with this. When younger I thought I could read the material and say "Oh, okay that makes sense I understand this." Only to fail miserably when on a test or somewhere I had to apply what I "knew" and realizing I didn't actually know it. I lean strongly toward autodidacticism and learned that if I could solve problems with the technique then I knew it.

My calculus 1 professor gave the advice that the best way to study is to do every problem in the book, then go and get another book and do every problem in that book, and keep doing that until you look at a problem and know exactly every step to do immediately.

Re: So you want to learn physics (2021)

#62

Post by kingkongjaffa »

carabiner wrote: post #23
kingkongjaffa wrote: post #13

It’s a decent list but thermodynamics is introduced way too late in the list (after modern and quantum wtf) and fluid dynamics and fluid mechanics including aerodynamics is entirely missing. If you have a physics education (I have an engineering education) can you tell me if you can really get a physics degree without bumping into Bernoulli or Navier-Stokes? At least just to see the lay of the land.

NS is too applied and seems akin to including the beam equation.

Would you also say Maxwell's equations are "too applied"?

They're both pretty firmly physics things rather than applied or engineering things, in their purest forms...

Re: So you want to learn physics (2021)

#63

Post by nologic01 »

The title should probably be: "So you want to learn theoretical physics".

While generally little known and appreciated among modern theorists and mathematical physicists, physics is actually an empirical science. In other words, every single section of that reading list is based directly or indirectly on a diverse and sophisticated set of devices and measurement configurations (aka experiments). Also, most progress in our understanding the physical universe follows simply from inventing ever better probes and opening new observation windows.

A computer analogy of the theoretical/empirical physics relation might be fun: You can spend your whole life writing application software and never even know what digital devices you are actually using. That's totally legit. But if you want to write a new computer language (= a new theory) you most likely will have to dig into memory architectures and caches and all that stuff. If you want to dramatically increase the speed of computation (= a new observation window) you have to design a new chip. And if you want to go really deep and invent new computing paradigms, well then you need to learn quantum mechanics :-)

In fairness, she does have a final sentence about that weird place called laboratory (= a place of labor).

> And, finally, a note on learning in a laboratory vs. learning from textbooks. Physics is both an experimental and theoretical science, and while research happens in laboratories and on blackboards and computers, the majority of any physics education does not take place in a laboratory but in lecture classes that teach from textbooks and assign homework problems that are found in textbooks.

My recommendation for a comprehensive intro into theoretical physics is The Road to Reality by Roger Penrose. Alas there is no such profound review of all experimental physics.

Re: So you want to learn physics (2021)

#64

Post by markus_zhang »

fcatalan wrote: post #10

I dropped out of Physics back in the day, because I loved computers a bit more, and now that I'm kind of fed up with computers I'd like to remove the thorn and do something like this. But I find that so much time has passed that I would need to brush up parts of my high school maths first, and this kind of discourages me before even starting.

I'm preparing to do something similar but attack a lesser beast that is the general relativity. I had a Master's degree in Statistics but unfortunately 1) Statistics does not really match Pure Mathematics and 2) I forgot most of it.

A beast it still is, I think it is contained in its own walls. I can skip any topic in Quantum Physics and others that is irrelevant.

I'm wondering if it's helpful to you too to focus on something smaller.

Re: So you want to learn physics (2021)

#65

Post by amluto »

This has the same omission that my undergrad program had: continuum mechanics. Even just the very basics (pressure, velocity, etc in a moving, non-equilibrium system) and translating between the terminology used by different science and engineering fields (static pressure, total pressure, velocity pressure, stagnation pressure, hydrostatic pressure, dynamic pressure, plain old pressure, head, oh my!) is very useful.

Hydraulics are everywhere. Ever used a sink? Flushed a toilet? Contemplated an air filter? Felt both sides of a small fan? Wondered how, exactly, a utility pump causes water to go in the inlet and out the outlet, and tried to read the manufacturer’s spec? Contemplated that the ripples when you throw a rock in an actual pond really don’t resemble the average “look I made water in WebGL” animation very much?

And more fancily, and very much in “Physics”, cosmological models usually model the universe as being full of a spatially varying continuous fluid. Stars are plasma or weirder things, and those are fancy fluids.

Yet, for some reason, the basics are missing from “Physics”. You can sometimes find them in mechanical engineering departments, and Feynman covers it a bit in his lectures.

Re: So you want to learn physics (2021)

#66

Post by brightlancer »

javajosh wrote: post #24
qsort wrote: post #3

A point that's rightfully emphasized by the author: > Solving problems is the only way to understand physics. There's no way around it. This generalizes well to other fields. I don't want to discourage anybody from trying to educate themselves in a difficult field (be it physics or something else), but that's a very common and immediately visible problem with autodidacts. If you haven't worked through enough hard pro…

Yes, and I'd take it further. What I'd like to see more of in physics texts is presenting a problem before offering the solution. Too often you get what amounts to a laundry list of techniques and ideas, which are the components to the answers to hard problems, but the student isn't motivated to learn them. If you present the hard problem first, the student may flail around and realize: I need something to help with…

> What I'd like to see more of in physics texts is presenting a problem before offering the solution.

Yes.

> If you present the hard problem first, the student may flail around and realize: I need something to help with this! Now that they know they need it, and why, you can give them the tool that fits the bill.

No.

If an instructor deliberately gives a student a problem that they know the student _cannot_ solve, then it rightfully destroys trust.

I never taught at the university level, but with middle and high school math students I taught them to how (re)discover the solutions, rather than teaching them the solutions directly.

As a practical matter, many of my college classes went too quickly to do anything _but_ teach the solution -- or tell us to learn it between classes and bring questions back.

Re: So you want to learn physics (2021)

#67

Post by monster_group »

computerfriend wrote: post #54
monster_group wrote: post #29

Earlier quoted context omitted.

I decided to start self learning theoretical physics late last year. I have been now studying physics every day for almost a year (before and after work). I did have to brush up on calculus and matrices but it came back very quickly (within a few days) after a 25 year gap so I'd say don't let that discourage you.

Excellent discipline. What topic are you focussing on?

I have been working on quantum physics since March of this year and am hoping to complete the whole text book (Townsend) by end of the year. Then on to special relativity -> classical field theory in 2024, general relativity and QFT in 2025 and 2026 - at least that's the plan.

Re: So you want to learn physics (2021)

#68

Post by BenFranklin100 »

ak_111 wrote: post #40
BenFranklin100 wrote: post #26

Earlier quoted context omitted.

I only became reasonably proficient in physics when I took the summer off between undergrad and graduate school and spent three months, six days a week, ten hours a day, doing nothing but working through four years of undergraduate physics curriculum by solving problems from my textbooks. There is no substitute for solving problems.

That is some dedication, well done. Interesting to know more about your career and where this lead to if you continued down the physics path.

I went on to earn a doctorate in biophysics and then began a career developing instrumentation for biomedical research. While I no longer work directly in the field of physics, the physical intuition and reasoning abilities honed by a physics education has allowed me to successfully lead teams consisting of electrical and mechanical engineers, while serving as the liaison to biologists and doctors. I regard a physics education as a modern-day ‘liberal arts’ technical degree.

Re: So you want to learn physics (2021)

#69

Post by froggit »

remote_phone wrote: post #46
qsort wrote: post #3

A point that's rightfully emphasized by the author: > Solving problems is the only way to understand physics. There's no way around it. This generalizes well to other fields. I don't want to discourage anybody from trying to educate themselves in a difficult field (be it physics or something else), but that's a very common and immediately visible problem with autodidacts. If you haven't worked through enough hard pro…

In college, I was always the first one out of my friends to “get” a concept, like fast Fourier transforms, anything with signal processing or even coding or any labs we had to do etc so I would spend time teaching them in the library. However I never did any of the exercises, mostly due to laziness and not arrogance. They would get A’s and I would get C’s and D’s. I emphasize this story to my kids because knowing isn…

This is kind of interesting to hear because I was the other way around. I found the best way to understand something was to teach it to others. That way I took what I already understood and was able to see what other people misunderstood, which was often something I'd never expected to be an issue, and add their experience in learning the topic on to what I already knew which expanded my overall understanding.

Then again, in the process of teaching I always found myself teaching people to work problems, which required me to be able to work the problems myself. In a way, it's kind of impressive you managed to avoid doing that.

Re: So you want to learn physics (2021)

#70

Post by growingkittens »

fcatalan wrote: post #10

I dropped out of Physics back in the day, because I loved computers a bit more, and now that I'm kind of fed up with computers I'd like to remove the thorn and do something like this. But I find that so much time has passed that I would need to brush up parts of my high school maths first, and this kind of discourages me before even starting.

Brushing up on high school maths is easy using Khan Academy, if that's all that's stopping you.
Post reply on HN