我遇到过一些作为枚举编写的状态机的例子。例如:
public enum State {
INITIAL {
@Override
public State process(char ch) { return MIDDLE; }
},
MIDDLE {
@Override
public State process(char ch) {
switch (ch) {
case 'a': return INITIAL;
default: return FINAL;
}
}
},
FINAL {
@Override
public State process(char ch) { return FINAL; }
};
public abstract State process(char ch);
public static State initial() { return INITIAL; }
}
让我感到担忧的是让常数彼此了解。 它们的实现可能会发生变化,如果添加(或删除)常量,则会隐含更多变化。
但是,除非外部代码直接引用常量并对转换表做出假设,否则所有更改都包含在枚举中。
那么,这段代码是高度耦合的,松散耦合的,还是主观的?
答案 0 :(得分:2)
我认为这对于灵活的工具来说太紧密了。
更宽松的东西会更灵活:
public enum State {
INITIAL,
MIDDLE,
FINAL;
public static State initial() {
return INITIAL;
}
}
// A 'null' for the Character matches anything.
static Map<Pair<State,Character>, State> machine = new HashMap<>();
static {
// Always transit from INITIAL to MIDDLE
machine.put(new Pair<State,Character>(State.INITIAL,null), State.MIDDLE);
// An 'a' transitrs us to FINAL
machine.put(new Pair<State,Character>(State.MIDDLE,'a'), State.FINAL);
// Anything else takes us back to INITIAL
machine.put(new Pair<State,Character>(State.MIDDLE,null), State.FINAL);
// Stay in FINAL once there.
machine.put(new Pair<State,Character>(State.FINAL,null), State.FINAL);
}