Hi David, I've started to compile a datatype for our molecule description that is also valid CML. We will only use a tiny subset of CML, but this way the Gromacs "general" topologies will in theory be readable by any program that can do CML. Cheers, Erik