In a derivative dY/dX, X and Y must have no axis names in common; it's up to the writer to rename the axes apart and to choose what the new names are. This means that there are no naming conventions (primes or stars) built-in to the formalism. (However, the naming convention used in examples is the same as other examples: primes for output axes.)
Differentials are used to get around the constant renaming that would have to be done if the chain rule were applied directly to complicated expressions.