Re: [PPL-devel] PPL's handling of floating-point coefficients