in practice it is sometimes a little more subtle than it seems.
e.g. we usually give a cyclic group in the form (Z/n,+) but may not notice that this gives not only the group but also a distinguished generator, namely 1.
the group( (Z/p)*, .) of units in the ring Z/p is also cyclic when p is a prime, but it may not be immediately clear what a generator is. moreover since there are many generators, (Z/p)* and no one is naturally distinguished, there is no completely natural way to choose an isomorphism between the groups (Z/p)* and Z/(p-1).
i.e. in a sense, ({1,2,3,...,p-1}, . ) are just different names for ({0,1,2,...,p-2},+) but it is not clear which new nAME CORRESPONDS TO WHICH OLD NAME.
e.g. when p = 7, (Z/7)* is isomorphic to Z/6, and the least generator is 3, perhaps it is natural to associate n with 3^n, for n=0,...,5, but I am not sure this generator is always best.