I'm fine with that. Just make sure that the conversion factor is not 1.0 ;) And tell me how that line in my example will look like, when you correct the bug.
Janek, This works as expected, to the best of my knowledge, for extracting conversion factors between units : template<class Y,class FromUnit,class ToUnit> inline Y conversion_factor(const FromUnit&,const ToUnit&) { return quantity<ToUnit,Y>(Y(1)*FromUnit()).value(); }; for which std::cout << conversion_factor<double>(CGS::dyne,SI::newton) << std::endl; gives 1e-05 If no named unit static const is available, you will, of course have to pass default constructed units : std::cout << conversion_factor<double>(CGS::momentum(),SI::momentum ()) << std::endl; I will commit this to the sandbox sometime this afternoon. Otherwise, for dimensionless heterogeneous quantities (mixed units of the same dimension), you should also be able to explicitly force the conversion via std::cout << quantity<SI::dimensionless>(1.0*CGS::dyne/SI::newton) << std::endl; At the moment this is still broken - we'll work on fixing it. I think that the conversion_factor<> syntax should be preferred anyway as it more clearly demonstrates the intent of the code. Cheers, Mattthias