What will be the number of states in minimal NFA that can be designed to accept the language $L = \left \{ \epsilon \right \}$ having a non empty alphabet set?
My friends are saying that this can be done using $1$ state, make the only state start state & final state without any transitions.
But will not this NFA accept every string over the alphabet set?
Because for every input string we are already in accepting state.
So every string will be accepted.
My claim is that at least $2$ states will be required so that we can get away from that default accepting state (start state) on any non empty string and the minimal NFA will be exactly same as the minimal DFA accepting $L$.
Am I correct?