1. A pleasant surprise
When I worked on the diameter problem for d-polytopes with n facets. I was aiming to prove an upper bound of the form 
It was a pleasant surprise to note that 
2. A bigger surprise
A few weeks ago James Lee gave a talk and proved a bound of the form 
(Update: I got it wrong at first, thanks guys)
This is an even more surprising special case of the formula above.
3. Is it better to have the discount first?
Question: What is a better deal: A store that gives 12% student-discount after it adds a 12% value added tax to the price of the product? Or a store that first adds 12% tax on the entire sum and only then deducts 12% student discount?
Ohh, The way I asked this question the two alternatives are precisely the same. Let me ask it again: What is a better deal: A store that gives 12% student discount after adding a 12% value added tax to the price of the product? Or a store that first deducts the 12% student discount, and only then adds 12% tax on the new price?
Answer: The same, by a surprisingly not obvious special case of commutativity of multiplication.
(See a related comment on Dave Bacon blog.)
Combinatorics and more 
Ok, so point (1) is explained by n^log d = e ^ ((log n) (log d)) = d ^ log n
But then, similarly for (2), how is
(log log log log n) * (log n) = (log log log n) * (log log n) ?
Am I missing something?
I would guess it should be Log[Log[n]]^(Log[Log[n]])=Log[n]^(Log[Log[Log[n]]])
What is not obvious about the third point?
Roland, I suspect this is much less obvious to most people than other cases regarding multiplication being commutative. This can be examined empirically, of course. (At the time, I found it confusing myself, but this may not tell much .)