Learning Tensor Basics: Overcoming Difficulties

[grzz]

I am learning about tensors. Can somebody give me some help. Thanks.

 

[DrGreg]

What you did was correct. You used<br /> \frac{\partial X^\mu}{\partial x^\alpha} \, \frac{\partial X^\nu}{\partial x^\beta} R_{\mu \nu} = r_{\alpha \beta}<br />and put r_{\alpha \beta} = 0 to prove R_{\mu \nu} = 0. You could have done it the other way round, i.e. put R_{\mu \nu} = 0 to prove r_{\alpha \beta} = 0. As we are talking about two arbitrary coordinate systems, that is equally valid.

 

[clamtrox]

You can kind of see this directly. If all of the tensor components are zero in one coordinate system, they must be zero in all others as well. So to show this, just show that the tensor itself is zero,

R = R_{\mu \nu} (dx^{\mu} \otimes dx^{\nu}) = 0

and since the tensor itself is a geometric, coordinate independent object, you're done.

 

[Chestermiller]

Here's another way of doing the problem. Write the starting equation out in component form. You will realize before you are finished writing that the starting equation represents a set of 16 linear homogeneous algebraic equations in 16 unknowns (the 16 components of R). Since the equations are homogeneous, the only solution to this set of equations is for each and every unknown to be zero.

chet

 

[grzz]

What you did was correct. You used<br /> \frac{\partial X^\mu}{\partial x^\alpha} \, \frac{\partial X^\nu}{\partial x^\beta} R_{\mu \nu} = r_{\alpha \beta}<br />and put r_{\alpha \beta} = 0 to prove R_{\mu \nu} = 0. You could have done it the other way round, i.e. put R_{\mu \nu} = 0 to prove r_{\alpha \beta} = 0. As we are talking about two arbitrary coordinate systems, that is equally valid.

So to remove \frac{∂X^{μ}}{∂x^{α}}\frac{∂X^{σ}}{∂x^{β}} from the lhs of \frac{∂X^{μ}}{∂x^{α}}\frac{∂X^{σ}}{∂x^{β}} R_{μσ} = 0, do I have to show all the steps as I did before (see the pdf in my origianal post) or can I just say that I divided both sides by \frac{∂X^{μ}}{∂x^{α}}\frac{∂X^{σ}}{∂x^{β}}?

Thanks everybody for all your help.

 

[Chestermiller]

If you have 16 linear algebraic equations in 16 unknowns, and the right hand sides of all the equations are equal to zero, what is the solution for the unknowns?