Archive for September, 2004

Geometry

Friday, September 10th, 2004

What exactly is geometry? I have yet to find a satisfactory definition. All the ones I’ve seen say it is something like the study of invariants under certain groups of transformations. Like Euclidean geometry is the study of invariants under the transformations of rigid motions. That I understand, but what about differential geometry, for instance? In that case, what would the relevant invariants and transformations be? Maybe arc length, curvature, torsion? And the things I don’t know anything about yet, like the various surface curvatures (intrinsic and extrinsic)? It isn’t really so hard to imagine the invariants, but what is the transformation group?

Anyhow, here’s a question I just came up with: Usually, we view rigid motions as reflections, translations, and rotations in Euclidean spaces. But, we can generalize the notion, by saying a map between two homeomorphic metric spaces is a rigid motion if it preserves distances. The question is simply to show that, when applied to Euclidean spaces, this definition of a rigid motion is equivalent to the standard one.

Greens Theorem

Sunday, September 5th, 2004

Green’ s theorem is something we never covered in my Cal III class. It never seemed much more than a curiosity to me until today; yet another piece of evidence that I should really have my calculus down cold before trying to venture into differential geometry. Today, I saw it being used to find the area enclosed by a parameterized simple closed curve. I’m going to apply that method to finding the area of an ellipse, and compare it to the method I would have used before.

Using Green’s theorem

Green’s theorem: \Displaystyle \iint_{{\rm int}(\vfun{\gamma})} \left(\frac{\partial g}{\partial x} - \frac{\partial f}{\partial y} \right) dx\; dy = \int_{\vfun{\gamma}} f(x,y) dx + g(x,y) dy, where f(x,y) and g(x,y) are smooth functions and  \vfun{\gamma} is a simple closed curve.

This gives an easy formula for the area enclosed by a simple curve,  A({\rm int}(\vfun{\gamma})) = \iint_{{\rm int}(\vfun{\gamma})} dx dy. Namely, let  \vfun{\gamma}(t) = (x(t), y(t)) . Then let  f = -1/2y and  g = 1/2x. Then by Green’s theorem,  A({\rm int}(\vfun{\gamma})) = 1/2 \int_{\vfun{\gamma}} x dy - y dx , so  A({\rm int}(\vfun{\gamma})) = 1/2 \int_0^a \left(x\dot{y} - y\dot{x}\right) dt , where a is the period of \vfun{\gamma}.

In the case of an ellipse  \vfun{\gamma}(t) = (a\cos(t), b\sin(t)) , this gives the simple expression  A = \frac{1}{2} \int_0^{2\pi} ab\cos^2(t) + ab \sin^2(t) dt = \pi ab!

Using naive integration and symmetry

First we find eliminate the parameter  t and find an expression for  y in terms of  x ,  \left(\frac{x}{a}\right)^2 + \left(\frac{y}{b}\right)^2 = 1 . Solve this for an expression for  y(x) as  x varies from  -a to  a :  y(x) = b\sqrt{1 - \left(\frac{x}{a}\right)^2} . Then notice that the area of the ellipse is twice that of  \int_{-a}^a y(x) dx , or  A = 2\int_{-a}^a b\sqrt{1 - \left(\frac{x}{a}\right)^2}  dx = 2b \int_{-a}^a \sqrt{1 - \left(\frac{x}{a}\right)^2}  dx. Recall that the antiderivative of  \sqrt{1 - \left(\frac{x}{a}\right)^2} is   \frac{1}{2} \left( x \sqrt{ 1 - \left(\frac{x}{a}\right)^2} + a\arcsin\left(\frac{x}{a}\right)\right) — or more likely, give up. Then plug and chug:  A = b \left[x \sqrt{ 1 - \left(\frac{x}{a}\right)^2} + a\arcsin\left(\frac{x}{a}\right)\right]_{-a}^a = ab \left( \arcsin(1) - \arcsin(-1) \right) = \pi ab

Obviously the second method is more error prone and requires more familiarity with your integral tables. Arguably, it is also less elegant than the first method. Also, the first method is easily extensible to more asymmetrical closed curves, whereas to use the second method would involve breaking up the region bounded by asymmetrical curves into several easier to handle curves. So, the first method is probably more practical to use overall. I wonder if there are any specific cases where the naive method is best?

Feynman’s Lectures on Physics

Saturday, September 4th, 2004

I’ve starting reading Feynman’s lectures on physics, the first volume. His style is refreshingly nonmathematical so far. However, it may be too nonmathematical– in the first serious chapter, his applications of the law of conservation of energy are done rather haphazardly– I’m sure the conclusions are correct, but not so sure that the method he purports to use is actually right. One of my profs (math), saw me reading it, and mentioned that Feynman is not very mathematically rigorous, and a friend of mine who is a physics major swears that at least one of Feynman’s derivations violates some physical laws. His specific example was that he derives the laws of homogenous wave propagation for sound waves. I don’t understand the problem yet, but all this naysaying is giving me pause.

Uploading Math notes

Friday, September 3rd, 2004

Finally, there is a good chance that my dream of uploading math notes to the site will become a reality some time soon. Since WP supports LaTeX through the LaTeXRender plugin, I don’t have to bother with writing my own script. The only thing I need do is make sure that all the custom commands and environments I used in my notes are defined, and to grep my TeX code for dollar signs and replace them with [tex] commands. I would like to hack LaTeXRender so that it handles displaymode and inline mode separately and appropriately, but that is too much crap to handle. I’ve tried modding it so that MetaPost can be used, but that’s not going to happen. I’m too lazy.

I can’t resist listing stuff, so here’s one of all the notes I’m aching to upload:

  • Differential Geometry notes: from the class I’m auditing this semester, and two other really good books.
  • Advanced Multivariable notes: from a class I took a couple of years ago.

I guess that’s a pretty short list… it makes me so proud of myself :)

Today I updated my quotes page, on my school site.