Hacker News Reader

Comment by opheliate | Hacker News Reader

Wow, these look great! Does anyone know of any similar resources for mechanics, specifically the Hamiltonian & Lagrangian formulations? I've had a bit of trouble finding good resources online to supplement my mechanics modules at uni.

View on HN

I found these lecture notes by David Tong to be really good at a glance [1]. A free introductory physics book [2]. I don't have much physics stuff and I know almost nothing about it, I mostly focus on Math/CS, this is some stuff that I had bookmarked. Maybe someone else here can share some good Mechanics resources.

[1] http://www.damtp.cam.ac.uk/user/tong/dynamics.html [2] http://www.lightandmatter.com

Based on "uni" I'm guessing you are at a UK institution (as am I).

The notes are utter dogshit because there's no culture of publishing them online for the public.

Kibble's classical mechanics book is deep, easy to read and cheap, I'd start there. Most physics books past a certain level end up with an increasingly hand-wavy derivation of the Euler-lagrange equations in some context

Tong is good, most appropriate if you already know a little about the subject. However, it is a "fast-track" kind of resource, it is good to read some standard textbooks on the subject, i.e. Landau & Lifshitz (the most efficient book on theoretical mechanics ever) and Goldstein (broad and interesting).

Although I have immense respect for them I really don't like L&L all that much. The first one is pretty beautiful but it could use some material on the "deeper" side of things, particularly the connection with geometry and Noether's theorem. As for the others, they're just too "russian" - I like how oddly practical they are, but they're just not structured properly - StatPhys in particular just doesn't flow very well.

The typesetting is also a crime against humanity - they should be done similarly to the Feynman lectures, so typed up in LaTeX rather than whatever they are now

Regarding geometry and Noether's theorem, I disagree. That stuff is way overblown in general, it gives you very little ( in classical mechanics) for the time and page count the subject requires. I had a lecture on this topic from enthusiast well-respected physicist, and I kept thinking, this stuff is interesting math but completely useless to a physicist.

Classical mechanics is ultimately just masturbation compared to the "real thing", it's all a warm-up for the mind rather than teaching you how to calculate stresses on a beam - Full on symplectic geometry is niche but the basics of Noether's theorem and symmetries are central to classical and quantum field theories which are ultimately the language in which the cutting edge of fundamental physics speaks.

So I sort of agree, but take the two Oxford books on QFT and Statistical physics - they pack in enormous amounts of detail in a 3 or 4 hundred pages. There's room to spare if the book is the right shape.

What are the titles of the two Oxford books you mentioned? I'd like to look at them.

Quantum Field Theory for the gifted amateur (well not quite...)

Concepts in thermal physics

Not at the same level as L&L but they are absolute masterpieces in pedagogy and the formula should be extended to a 800page tome to cover more material. It's the way MTW should be written...

> could use some material on the "deeper" side of things

The first few chapters are precisely about the extraction of concrete conserved quantities from symmetries. A sufficiently "russian" reader will be able to see immediately the general case from these very complete examples. Noether's theorem is only missing in name, not in content.

> The typesetting is also a crime against humanity - they should be done similarly to the Feynman lectures, so typed up in LaTeX rather than whatever they are now

The typeseting of the french edition of L&L is extraordinarily beautiful. I'm not sure that is is possible to reach this typographical elegance with freely available LaTeX packages. On the other hand, the re-typed Feynman lectures are a disgrace, and an objectively worsening in quality from the first editions.

I liked that section a lot but I want a jumping off point. I don't think symmetries of the action or groups are discussed, IIRC.

I may have been a little vague when I said typesetting - the actual typesetting is a bit dull but the thing that is a disgrace, I should clarify, is that the fonts are terrible, and the text has that "Cheap Dover book" print quality (not far off a good typewriter).

To amplify this point, all the negative Amazon reviews praise the book but excoriate the print quality.

This must concern some paperback English print? My copy in French from "Editions MIR Moscou, 1969", hardcover, masterfully typeset and in bible paper is one of the finest books that I have.

Yes, I think it's specific to the English version. You can see examples if you use the Amazon.com "look inside" feature. For instance in volume 3, in the first pages already some of the footnotes are hardly readable.

For one of the volumes, a reviewer even posted a picture of his edition with subscripts printed so faint/partially that they were basically missing, and he had to Google the proper formulas!

Yes, I literally have RF engineering books (interesting reads actually) that are as old as my grandfather that have clean equations and beautiful text, in a lovely binding, but my L&L copies are both like an n-th order photocopy.

Yeah, I liked the mechanics and quantum volumes, but the choice of material in the stat mech books (both volumes) just seemed weird compared to what's usually done in a modern graduate course.

Do you have any recommendations for good stat mech books? I guess the two volume set by Kardar is trendy right now, but I'm always curious about other options.

I'm remiss to recommend as I'm still learning, but I absolutely love "Concepts in Thermal physics". It's embarrassingly well written and structured.

Beyond that the usual suspects are good enough for me, I just don't like L&L. I'm probably forgetting something.

It's not exactly statmech but I really like McComb's introductory renormalization methods book.

Thank you for the recommendations!

While I have not seen either of those books before, I have another book by the author of the first one, Stephen Blundell. It's Quantum Field Theory for the Gifted Amateur, and I think you might like it. Not as a standalone quantum book, but perhaps as a supplement to a more "serious" treatment.

edit: I just saw you recommended this below!

QFTFTGA is really great. Surprisingly hard as well.

I got it when I was 17 and it was an absolute godsend as its really cheap.

Look for Dover books, particularly from Russian authors. Cheap and good!

For linear algebra: Shilov.

I own four copies of that book :-)

Two hard copies since I moved countries, one on kindle and one on Apple Books.

Great book :D

Reply on news.ycombinator.com