By an "invariance under reformulation" argument, show that a reasonable noninformative prior for a Poisson parameter f) is   random variable X usually arises as the number of occurrences of a (rare)...

By an "invariance under reformulation" argument, show that a reasonable noninformative prior for a Poisson parameter f) is
  random variable X usually arises as the number of occurrences of a (rare) event in a time interval T. The parameter f) is the average number of occurrences in time T. Since the specification of T is rather arbitrary, consider the experiinent which would result if the time interval used was cT (c > 0). This idea was due to Jaynes (1968).)