home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!caen!sol.ctr.columbia.edu!shire.math.columbia.edu!dy
- From: dy@shire.math.columbia.edu (Deane Yang)
- Subject: Re: What's a manifold?
- Sender: nobody@ctr.columbia.edu
- Organization: Mathematics Department, Columbia University
- Date: Mon, 9 Nov 1992 19:34:45 GMT
- Message-ID: <1992Nov9.193445.15617@sol.ctr.columbia.edu>
- X-Posted-From: shire.math.columbia.edu
- NNTP-Posting-Host: sol.ctr.columbia.edu
- Lines: 62
-
- As a practicing differential geometer, the atlas definition of
- a manifold is the most "natural" (in the long run) and the most
- useful. Baez has argued for the scheme definition. This, in
- general, has proved less useful, except for those doing
- algebraic geometry and those doing Alain Connes' noncommutative
- differential geometry (These are both very important areas
- of current research).
-
- The problem in teaching this definition, however, is that I also
- believe that when you're introducing a new concept to students,
- it's important to be completely precise and unambiguous in
- your notation, so that students have less to get confused about.
- But all this discussion about atlases and maximal atlases
- plus all the indices up and down labelling the open sets and the
- coordinates makes for a big mess that obscures the essential
- ideas.
-
- If I could control how students learn about manifolds from the
- start, I think I would follow history (as was described in
- an earlier posting). Teach the students about 2-manifolds
- and Riemann surfaces first. Riemann surfaces can be motivated
- very nicely using complex analysis. Moreover, I think it's
- much easier to work with a holomorphic manifold at the start.
- Also, since the dimension is 1, there's one less index to
- clutter up the notation.
- From there, it's not hard to generalize to higher dimensions
- and less regularity.
-
- I'm teaching an introduction to manifolds course right now.
- What I chose to do is to first define the notation of a
- "local manifold", which is simply a manifold that is
- diffeomorphic to an open set of Euclidean space.
- I have emphasized an analogy with abstract linear algebra,
- where a vector space is really the same thing as R^n, but
- where we've chosen treat all bases on equal footing.
- A local manifold is an open set in R^n, where we treat
- all possible coordinate systems on equal footing.
-
- The advantage I see in doing this is that I only
- need one coordinate chart at a time and I don't have
- to mess around with covers of charts. I'm now developing
- the usual foundations of manifolds in this limited setting.
-
- Later, I will observe that local manifolds can be
- glued together in the obvious way, and we'll get
- the standard notion of a manifold. I'll also
- discuss why everything we've done so far generalizes
- to manifolds.
-
- When teaching the "local theory of manifolds", which is
- what is usually done in an introductory course, I always
- find myself emphasizing that the notion of a manifold
- is in some sense a nonlinear analogue of the notion of
- a vector space.
-
- The problem with this approach is that it assumes the
- students have come to terms with abstract linear algebra.
-
- Any suggestions and ideas about what should be taught
- in such a course and how the course should be taught
- are welcome.
-
-