In both XML and ChemKin formats, the NASA thermo data (coefficients) and transport data are listed by species. Perhaps it makes sense to cache those data (using existing Antioch objects) within the ChemicalSpecies? It would make parsing much easier (because it's done at the time when we parse the species data) and then deduction of the thermo type is easier (we just query the chemical species instead of trying to poke in the input file or make assumptions based on the file type).