What should have happened if Hardy had discovered this? (Q368028): Difference between revisions

In this excellent paper the authors present a simple proof of Hardy's inequality \[ \int_0^{\infty}\left(x^{-1}\int_0^x f(y)dy\right)^p dx \leq \left(\frac p{p-1}\right)^p\int_0^{\infty}f^p(x) dx, \] by remarking that it is equivalent to the inequality \[ \int_0^{\infty}\left(x^{-1}\int_0^x g(y)dy\right)^p \frac{dx}{x} \leq \int_0^{\infty}g^p(x) \frac{dx}{x}, \] and using Jensen's inequality and a reverse in the order of integration. They then ask and investigate what might have happened if this proof had been discovered by any of the various mathematicians involved in the original discovery and proof of the inequality. They use this method to obtain a generalisation of Hardy's inequality for finite intervals with sharp constants and to obtain one with powers \( p= p(x)\) that are piecewise constant.