Klockan3 said:
n physics graduate courses you often go through more maths than in a mathematics graduate course, except that you don't do it thoroughly in the physics one.
Well, yes and no. The topic of the courses are really very different and only superficially related. It's like the difference between introductory calculus courses, and a real analysis course -- although superficially the same subject, the topics of the courses are actually very different.
The calculus class is more geared to how to do calculations with derivatives, integrals, series, and approximation methods, and how to use these tools to solve problems.
The real analysis class, on the other hand, is more focused on how to
build the tools of calculus rather than how to use them. Real analysis has its own collection of tools and techniques that are useful to that purpose, and can be used to construct new sets of tools in novel situations.
However, in texts where it is appropriate, it is common to see things like:
Lemma: blah blah blah
Proof: See [some other text]
and then that lemma is subsequently used in the next argument. Sometimes you even see appendices that boil down to things like "homotopy theory in five pages!" Such a thing would:
- Give names to the objects of homotopy theory that will be used
- State how to manipulate those objects. (often in theory form)
- State a few proofs / exercises to give the flavor of the techniques used to manipulate those objects to a fruitful purpose
- Give references for further reading
People who studies mostly maths gets too reliant on that rigor to be able to study physics. I Mathematicians aren't used to this, they can't keep up using only the skills they developed at the maths department.
My last point above, I think, is the key difference. (admittedly, I'm now speaking on little information)
A physics student might be trained to see new manipulations and learn how to repeat them, and to ignore things that don't quite add up. A mathematician, however, brings a different skill-set and is trained to be creative, and is likely to run into the things that don't work more quickly and become frustrated trying to wade through the white lies to see the truth. (e.g. treating position eigenstates as if they were actually quantum states)
For a personal anecdote, I can't express how irritated I was when I finally divined that most integrals and limits that physicists write are meant in the distributional sense, which boils down to being meant to be evaluated in the reverse of the order there written.