Andrej Bauer wrote: > it is non-Cauchy in a strong sense: for every epsilon > 0 and k > there is m such that |a_k - a_n| > epsilon for all n > m. It seems impossible that we can take epsilon > 1. GL