Comment on `Asymptotic Scaling in the Two-Dimensional O(3) sigma-Model at Correlation Length 10^5' by S. Caracciolo et al

Abstract

We explain why in our view an extrapolation from small lattices containing only perturbative information cannot be sufficient to determine nonperturbative qunatities and therefore cannot lead to a trustworthy determination of the correlation length.

Comments: 3 pages, latex, no figures