Sometimes its not obvious what sould go into a separate step and what to pack directly into a sequence. Let's study the following example.
A robot finished a homing sequence. It should move to a ready position before further action can happen. For this purpose a sequence readying is implemented as follows:

class Readying : public Sequence {
public:
  Readying(std::string name, Sequencer& seq) : Sequence(name, seq) {setNonBlocking();}
 
  int action() {
    cs.pathPlanner.move(readyPos);  // we assume that the control system comprises of a path planner
  }
 
  bool checkExitCondition() {
    bool end = cs->pathPlanner.endReached();
    if (end) safetySystem->triggerEvent(safetyProperties->readyDone);
    return end;
  }
};

The sequence consists of a single action - set the final destination of the path planner. It will terminate as soon as the path planner has reached this position. Before returning it will trigger a safety event which causes the safety system to switch to the next level.

The same goal as above could be achieved as follows:

class Move : public Step {
  Move(std::string name, Sequencer& seq, BaseSequence* caller) : Step(name, seq, caller) { }
 
  int action() {
    cs.pathPlanner.move(readyPos);  // we assume that the control system comprises of a path planner
  }
 
  bool checkExitCondition() {
    return cs->pathPlanner.endReached();
  }
}
 
class Readying : public Sequence {
public:
  Readying(std::string name, Sequencer& seq) : Sequence(name, seq), move("move", seq, this) {setNonBlocking();}
 
  int action() {
    move(readyPos);
    safetySystem->triggerEvent(safetyProperties->readyDone);
  }
private:
  Move move;
};