# String theory question

1. Sep 14, 2007

### ehrenfest

1. The problem statement, all variables and given/known data
I saw the following equation in my (Zwiebach page 83).

$$2 ds \delta (ds) = \delta (ds) ^2$$

where delta is the variation from variational calculus and ds is the Lorentz invariant spacetime distance.

It seems like they took a derivative from the right to the left but I am really not sure why you can do that because I thought delta was just a very small variation function.

2. Relevant equations

3. The attempt at a solution

2. Sep 14, 2007

### BoTemp

The delta notation means "change of" which is the same as derivative. If it makes you feel better, replace delta with delta/ dx and then multiply by dx on both sides to remove it. Seems a bit sketchy, but it works as long as the variation is small, which of course is the point of derivatives in the first place.

3. Sep 14, 2007

### ehrenfest

So, in this case it makes the most sense to replace delta by d/ds, right? That does seem sketchy.

Last edited: Sep 14, 2007
4. Sep 15, 2007

### Dick

It's not sketchy. d and delta are two different things. d refers to the differential of s(t) with respect to a parameter as in s(t)->s(t+epsilon). delta refers to the variation of some functional with respect to the variation of s(t) by a arbitrary function. s(t)->s(t)+epsilon*r(t).

5. Sep 16, 2007

### Jimmy Snyder

As Dick wrote, d and $\delta$ are not the same thing. ds as it is said, is the ghost of a departed quantity, while $\delta s$ is a boring old number.
$2dsd(ds) = d(ds)^{2}$ is an exact equation.
$2 ds \delta (ds) = \delta (ds) ^2$ is approximate. To be exact, it should be
$2ds \delta(ds) + \mathcal{O}((\delta ds)^2) = \delta (ds) ^2$
In this case, you are expected and indeed required to use only such $\delta$ as to make $\delta^2$ small enough to ignore.

Last edited: Sep 16, 2007
6. Sep 29, 2007

### ehrenfest

Sorry--I encountered this again and I am still confused about it. When you say replace delta by delta/ds (and then multiply both sides by ds), do you mean that I should treat this as the derivative operator with respect to ds? But can you do that if there is more than just one term in there (say $$\delta (dx dy dz) ^2$$)? How do you think of it then?