Actually, I would not recommend using any rules unless you really understand why and how they work. I.e. you should be able to derive those rules from first principles. Otherwise you are just training to become some sort of math computer who does not really understand math :)
L'Hopital's rule is actually closely related to Taylor's theorem which says that foir a sufficiently differentiable function f one has:
f(x+h) = f(x) + h f'(x) + h^2/2 f''(x) + h^3/6 f'''(x) + higher order terms in h.
This is something that you can easily derive and intuitively understand. Near a point x the function f will be f(x) plus correction terms. You can apply this to compute the limits as follows:
In case of (lnx)^(x-1) you take the logarithm, using that the log of the limit is the limit of the log which follows from the fact the the logarithm is a continuous function.
You then get the function
of which you want to compute the limit x--->1
You then put x = 1 + h and do a series expansion in powers of h. You get:
ln(1 + h) = h - h^2/2 + O(h^3)
where O(h^n) means a term proportional to h^n for small h. Taking the log again and using the same series expansion gives:
ln(ln(1 + h)) =
ln(h - h^2/2 + O(h^3)) =
ln(h) + ln(1 - h/2 +O(h^2)) =
ln(h) - h/2 + O(h^2)
Multiplying this by h and taking the limit h -->0 gives -1/2, so the original limit is exp(-1/2)
Note that we needed to go to second order in the Taylor expansion, which means that if you had used L'Hopital you would have had to use that rule twice.
Sorry, I was a bit confused, when multiplying by h the limit becomes zero so after exponentiationg you get 1. So, You could have used L'Hopital's rule once...
im going to figure this one out!