Subject: Handling branch cuts for hypergeometric functions
From: Raymond Toy
Date: Fri, 25 Feb 2005 09:22:34 -0500
>>>>> "Wolfgang" == Wolfgang Jenkner writes:
Wolfgang> Raymond Toy writes:
>> These kinds of things always reminds of an example given in Kahan's
>> paper Much Ado about Nothing's Sign. He has an example two fairly
>> complicated functions and asks if they're equal. If you plug some
>> random values, they are equal. I think that if you do some naive
>> manipulations, you'll prove that they are equal. In fact, they are
>> equal, except for in a small oval region of the plane, which you'd
>> probably not find by random testing. The difference arises because of
>> the branch cuts of the functions involved.
Wolfgang> But the `typical values' select only the premises of an implication,
Wolfgang> say
Wolfgang> f=g => some result
Wolfgang> It's still up to the user to work out if f=g is satisfied or not.
Wolfgang> At least this is how I understand Albert Reiner's proposal in this
Wolfgang> thread.
I was just wondering how useful it would be in more complicated
situations. The example I gave was a great example but I think it
illustrates the pitfall of using "typical" values to decide
something. But perhaps for the kinds of questions maxima asks, it is
ok. I certainly haven't come up with a counterexample where this
would not work.
Ray