Feynman on indefinite integral?

75 Views Asked by At

Consider:

$$ \int \frac{ \ln (t-1)}{t} dt$$

And then,

$$ I( \alpha) = \int \frac{ \ln [\alpha (t-1)] }{t} dt$$

D.w.r.t. $\alpha$

$$ \frac{dI}{d \alpha} = \int \frac{ 1}{ \alpha t} dt$$

$$ \frac{dI}{d \alpha} = \frac{1}{\alpha} \ln t +C \tag{1}$$

Integrating both side with $\alpha$,

$$ I = \ln \alpha \ln t + C \alpha$$

And putting $ \alpha =1$

I get a weird result...:

$$ I = C$$

What did I do wrong..? Also is the constant in (1) a function of $ \alpha$ ..? (generally speaking)

1

There are 1 best solutions below

0
On BEST ANSWER

Ok, the differentiation is where I messed up.

$$ I_{\alpha} = \int \frac{\partial}{\partial \alpha} \frac{ \ln (\alpha (t-1) )}{t} dt$$

Or,

$$I_{\alpha} = \int \frac{1}{t} dt$$

Hence,

$$ I = \ln t \alpha +C$$

Hence,

$$ \int \frac{ \ln [\alpha (t-1)] }{t} dt = \alpha \ln t + C$$

Too bad there is no easy points to evaluate this at >:(