Physics Grad Programs that don't require Jackson's Electrodynamics

[Prologue]

Jackson suits a certain type of student very well. If you have a good, careful, thorough professor, like to read a book thoroughly (~20hours a week) and then straight up copy the methods out of the book when you are solving the exercises, you will do fine. If you are a more 'intuitive' person that would like at least a glancing blow of an explanation on things, would like to sit down, read the problem, think about it and then answer it, good luck. Also, if you don't like to wade through ~20 hours of (excruciating) reading a week...you will just have to pray/hold out longer than the other half of the people in the class.

Also, there is a major difference between working 60hours a week on something that is interesting and working/reading Jackson for 30+hours, and using the other 30 hours on classes/colloquia/other courses. Jackson sucks the life out of you and pulls you away from your real interests. Of course, maybe the skills gained will be worth it in the end. I doubt it, but can't say. One thing that is puzzling to me in the asymmetry between the two classical topics, EandM and Mechanics. Nearly every school requires a 2 semester Jackson course, but some don't even require one semester of a theoretical mechanics course. That is just offensive in my opinion. How can EM be that important?

 

[plasmaqueen]

I have spent 8 months studying in Jackson's first 6 chapters. It is a masterpiece, I have to confess. It requires a strong background in mathematical physics techniques. The trick is to fully understand the theory, and to handle solutions by Separation of Variables really well. It boils down to finding coefficients by applying boundary conditions.
Yet I agree that spending 50+ hrs reading through Jackson makes students no longer interested. But, here's the challenge, who is strong enough to go on?

 

[cgk]

I tend to be a traditionalist when it comes to physics education. One problem with trying to keep physics "relevant" is that people in academia often seriously guess wrong what skills are useful in industry.

But doing a wild guess might still be better than blindly following 100 year old leads.
I'm not talking about E&M specifically, but there is lots of new information coming up, and space needs to be found for it. This can only be done by removing old topics from the curriculum and putting in new ones.

For example, there is a four volume series of books by Felix Klein and Arnold Sommerfeld... on the theory of the spinning top. They go through all possible and impossible variations and solution attempts at the spinning top problem. And I imagine 100 years ago, when this was still hot stuff, physicists argued about whether or not every physicist should have a profound understanding of solution strategies for such rigid body motions... Now I know very little about the spinning top, and I don't think I missed that.


I think that updating the curriculum can be a good idea. For example, in my opinion almost all statistical mechanics textbooks and many many-body theory textbooks are very bad because they set the focus completely wrong due to following "traditional" approaches and topics instead of teaching what is actually relevant. All that is taking away space from the more relevant problems. E.g., in modern E&M, these would certainly include *numerical* solutions to the PDEs. Where these even covered in Jackson to any significant degree? I can't remember.

 

[chill_factor]

But doing a wild guess might still be better than blindly following 100 year old leads.
I'm not talking about E&M specifically, but there is lots of new information coming up, and space needs to be found for it. This can only be done by removing old topics from the curriculum and putting in new ones.

For example, there is a four volume series of books by Felix Klein and Arnold Sommerfeld... on the theory of the spinning top. They go through all possible and impossible variations and solution attempts at the spinning top problem. And I imagine 100 years ago, when this was still hot stuff, physicists argued about whether or not every physicist should have a profound understanding of solution strategies for such rigid body motions... Now I know very little about the spinning top, and I don't think I missed that.I think that updating the curriculum can be a good idea. For example, in my opinion almost all statistical mechanics textbooks and many many-body theory textbooks are very bad because they set the focus completely wrong due to following "traditional" approaches and topics instead of teaching what is actually relevant. All that is taking away space from the more relevant problems. E.g., in modern E&M, these would certainly include *numerical* solutions to the PDEs. Where these even covered in Jackson to any significant degree? I can't remember.

i agree. For what it's worth, my M.S. Physics program right now doesn't use Jackson and if I go for a PhD, the use of Jackson is going to be a selection factor.

i truly believe that very few people need to know how to analytically solve EM problems at the jackson level. like you said, today its going to be numerical.

 

[plasmaqueen]

Are you sure guys that we do not need to understand EM problems at the Jackson Level, I am at MS program that Jackson is a MUST. You see, I am starting to think that my thinking level has risen after the attempt of solving those problems.

As for numerical, definitely , we need to study EM in a numerical approach, is there any book that does this today?

 

[plasmaqueen]

Yes I agree about statistical mechanics, which book did you use?

I am very interested in those volumes about the spinning top, we spent many hours discussing it in class, i even brought a top home! We were referring to Goldstein's book. I even have painted the spinning top...in charcoal ;)

 

[chill_factor]

Are you sure guys that we do not need to understand EM problems at the Jackson Level, I am at MS program that Jackson is a MUST. You see, I am starting to think that my thinking level has risen after the attempt of solving those problems.

As for numerical, definitely , we need to study EM in a numerical approach, is there any book that does this today?

at least at my school, we don't.

for plasma physics, EM is very very important, but how much EM do you need to know to do condensed matter physics? there's chemists and materials engineers doing the same research and they literally know ZERO EM.

 

[v0rtex]

I have nothing helpful to say, but I just don't get it. I want to go to a grad school that actually HAS Jackson's electrodynamics as a requirement. IMO this is has to be the most interesting part of classical physics (the most interesting physics subject IMO).

Up until last year, my school had a mandatory 3rd year undergrad course in electrodynamics that followed Jackson's text and I wish I could have taken it, because I really feel that I haven't learned enough with my first year courses, 2 EM courses + an EM optics course. Here's a link to the course material, including a solved quiz/exam, in case anyone cares:
http://webpages.ull.es/users/ajhernan/electrodinamica.htm

 

[twofish-quant]

i truly believe that very few people need to know how to analytically solve EM problems at the jackson level. like you said, today its going to be numerical.

But I don't see how you can come up with decent numerical results without having some analytic grasp on the problem, and most numerical computation involves a large amount of analytical manipulation to get the problem in a form that you can do numerical stuff on it.

 

[twofish-quant]

This can only be done by removing old topics from the curriculum and putting in new ones.

What happens in research is that the new stuff you figure out, and you can figure it out by seeing how stuff was figured out before.

They go through all possible and impossible variations and solution attempts at the spinning top problem. And I imagine 100 years ago, when this was still hot stuff, physicists argued about whether or not every physicist should have a profound understanding of solution strategies for such rigid body motions... Now I know very little about the spinning top, and I don't think I missed that.

What I've seen in industry is that you'd be asked to develop an extremely deep knowledge on an obscure topic, and the goal of graduate research as I see it is to give you the skills so that you can develop this sort of deep knowledge very quickly.

E.g., in modern E&M, these would certainly include *numerical* solutions to the PDEs. Where these even covered in Jackson to any significant degree? I can't remember.

If you know Jackson you can spend two weeks reading numerical recipes and code up something that more or less works for that class of problems.

I agree that there is always too much to teach, but my philosophy is that if by knowing X, you can figure out Y quickly, then there is no real point in spending too much time on Y. No you don't know finite element methods, but I think that it is the case that if you know Jackson really well, you can figure out FEM quickly, but the reverse is not true.

 

[twofish-quant]

Nearly every school requires a 2 semester Jackson course, but some don't even require one semester of a theoretical mechanics course. That is just offensive in my opinion. How can EM be that important?

Classical EM is a solved problem. Classical mechanics isn't. You can spend two semesters and learn "everything that's worth knowing' about classical EM, whereas you can spend decades on one aspect of classical mechanics and still not get to the bottom of it.

One analogy is that if you learn painting, a lot of teachers will have you spend inordinate amounts of times painting a sphere or a cube. Once you have that, there are some standard subjects that you'll be asked to paint (i.e. fruit baskets).

Jackson is the physics equivalent of painting a cube.

 

[SophusLies]

Classical EM is a solved problem. Classical mechanics isn't.

I definitely agree but there are still surprises out there. One of my professors showed us this article a couple days ago:

http://physics.aps.org/featured-article-pdf/10.1103/PhysRevLett.108.163901

I am also in agreement of your learning philosophy. I've always been the type that wants to learn something as deep as I can in hopes that it will pay off somewhere else. In most cases, it has paid off even in surprising ways (representation theory, Lie stuff) but in others not so much (Galois theory, number theory..). Maybe it's helped my learning abilities but sometimes it doesn't feel like it.

On another note, I loved Jackson's book mostly for the real world nasty non symmetrical problems. It made me feel like I was getting close to model an actual physical phenomenon rather than a "thought" experiment one.