There's an old conjecture that, if you take a positive integer, and repeatedly reverse the digits and add the two mirror image numbers, you eventually get a palindrome. An article in Scientific American many years ago said this conjecture had never been proved true in any base, but had been proved false in the binary; the article gave the first known binary counterexample and sketched the proof that it was such. The first possible counterexample in base ten was given as 196, for which the process had been carried through tens of thousands of iterations without encountering a palindrome. Someone just informed me by E-mail that he was running the process for 196 on his computer, and had gone through over 200,000 iterations without finding a palindrome. If anyone has some references to this problem more recent than the article mentioned above [which I think is about 20 years old], I'd like to know so my correspondent's computer can take a breather!