I recognize myself in these notes by Chris Tiee on tensor analysis—the fever of acquisition, the uses of comparison, and the stray insight that makes the whole book worth it (as the one track does for the whole album):
“One way to learn a lot of mathematics is by reading the first chapters of many books.”—Paul R. Halmos
Ah, the dreaded discussion of texts for tensor analysis. I am addicted to collecting math books (I also often have 10 books checked out from the library simultaneously) and reading the first 20 pages of them. Very occasionally I make it through farther than that. Despite the proliferation of bad tensor analysis texts (some would say all of them are bad), I have to admit I have gleaned everything I have learned about Tensor Analysis from reading these books, collecting the knowledge into a gigantic patchwork. The fact is, each text actually has a gem of insight or two that is not presented in any others. There is much overlap in the bad parts, and some in the good parts too, but of course, it’s always hard to consult so many references, since I often forget the transformation laws on those overlaps. . . not to mention also the transformation laws that tell how the notation changes—regarding this, we have the following
A.1. Joke. Differential geometry is the study of those things invariant under change of notation.
Another problem is also that it’s very hard to strike a balance between being intuitive in the explanations of what these things are—and hence being vague—and also developing a precise, formal theory that is the real deal—hence being obfuscatory.
The quote from Halmos isn’t quite right. It’s actually, “I wish I had read the first ten pages of many more books—a splendid mathematical education can be acquired that way.” As for tensors, a newer book rich in intuition is Dwight Neuenschwander’s Tensor Calculus for Physics.