Mathematics desk
< June 17	<< May \| June \| Jul >>	Current desk >

Welcome to the Wikipedia Mathematics Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

June 18

Tautological one-form

I'm self-studying The Geometry of Physics by Frankel. I'm struggling to understand the tautological one-form, in particular trying to reconcile what Frankel says about it with what our article says about it. If I'm understanding Frankel right (which is a big if), it sounds like a tautological one-form is just the pullback of the projection from a cotangent bundle to the base manifold, or using $\theta$ and $\pi$ as defined in the article, $\theta =\pi ^{*}$ . At the very least, unless I'm confused, the two maps are between the right spaces, $\theta :M\to T^{*}M$ and $\pi ^{*}:M\to T^{*}M$ , where as defined in the article, $M=T^{*}Q$ . But both the "Coordinate-free definition" and "Properties" sections of the article make it sound like $\theta$ is more complicated than that. Is the article not giving the simplest explanation of what $\theta$ is, or is $\theta =\pi ^{*}$ incorrect? If $\theta =\pi ^{*}$ is incorrect, can you give an example that illustrates how that expression is sometimes or always wrong, or otherwise help me understand how I'm confused? Red Act (talk) 08:50, 18 June 2014 (UTC)[reply]

The trouble with this approach is that

\pi ^{*}

is not a function from

M\to T^{*}M

(nor is

\theta

, but for different reasons). It is also not true that

\theta =\pi ^{*}

. Here is a correct approach. A one form

\alpha

on M also defines a section of the cotangent bundle,

s_{\alpha }:M\to T^{*}M

. The tautological one-form is the unique one-form

\theta

on

T^{*}M

such that

s_{\alpha }^{*}\theta =\alpha

for all one-forms

\alpha

on M. Sławomir Biały (talk) 10:59, 18 June 2014 (UTC)[reply]

It appears that you and the article are using

M

to mean two different things, so I can't tell how much of what you said I was wrong about is just due to the notational difference, versus how much is due to my being confused. You're using

M

to mean the base manifold. The article uses

Q

to mean the base manifold, and

M

to mean its cotangent bundle,

M=T^{*}Q

. Also, for what you're denoting as

s_{\alpha }

, the article would just go ahead and denote as

\alpha

(see the first expression in the Properties section), in what I'm guessing you would consider to be an abuse of notation. Which things, if any, remain wrong in my post above, given that it uses the notation that the article uses, instead of the notation that you used? Red Act (talk) 15:43, 18 June 2014 (UTC)[reply]

Ahh... Your post makes sense now. Yes, the canonical one-form is just the section of $T^{*}M$ over M given by $\pi ^{*}$ . Sławomir Biały (talk) 17:39, 18 June 2014 (UTC)[reply]

Help with fixing some PARI/GP code

I've written the following PARI code:

c=0; forprime(a=2, 20, while(c<2, forprime(p=2, 10^9, if(Mod(a, p^2)^(p-1)==1 && c==0, c++, if(Mod(a, p^2)^(p-1)==1 && c==1, print1(p, ", ") && c++, if(Mod(a, p^2)^(p-1)!=1 && c==2, p=10^9 && c==0))))))

This is supposed to find the second Wieferich prime for increasing bases. It correctly reports 3511, but then there is no further output (a computation seems to happen, as the fan of my computer runs). I can't believe it should take very long to find 1006003, the second base 3 Wieferich prime. Is there an error in my code? -- Toshio Yamaguchi 10:29, 18 June 2014 (UTC)[reply]

There are several errors. For starters, you appear to use && as a statement separator. && means and, but print returns void (0 when used in &&), so you don't perform the statement after a print. The statement separator in PARI/GP is a semicolon. Maybe you have seen scripts which use && as a way to make multiple statements when the programmer knows each one, except possibly the last, returns non-0. You also have other errors and unnecessary repetition of the modular exponentiation. Here is how I would code it:

forprime(a=2, 20, c=0; forprime(p=2, 10^9, if(Mod(a, p^2)^(p-1)==1, c++; if(c==2, print1(p, ", "); break))))

Note that oeis:A178871(5) is unknown. PrimeHunter (talk) 11:08, 18 June 2014 (UTC)[reply]

@PrimeHunter: Thank you. I sent you an E-Mail. -- Toshio Yamaguchi 13:13, 18 June 2014 (UTC)[reply]