我如何参考Isar当前的子目标?

时间:2017-03-03 15:49:31

标签: isabelle isar

我正试图从Programming and Proving in Isabelle解决练习4.7。 我遇到了一个我证明是错误的情况,因此我发现了一切,但我无法结案,因为我不知道如何参考我的证明义务。

theory ProgProveEx47
  imports Main
begin

datatype alpha = a | b | c

inductive S :: "alpha list ⇒ bool" where
  Nil: "S []" |
  Grow: "S xs ⟹ S ([a]@xs@[b])" |
  Append: "S xs ⟹ S ys ⟹ S (xs@ys)"

fun balanced :: "nat ⇒ alpha list ⇒ bool" where
  "balanced 0 [] = True" |
  "balanced (Suc n) (b#xs) = balanced n xs" |
  "balanced n (a#xs) = balanced (Suc n) xs" |
  "balanced _ _ = False"

lemma
  fixes n xs
  assumes b: "balanced n xs"
  shows "S (replicate n a @ xs)"
proof -
  from b show ?thesis
  proof (induction xs)
    case Nil
    hence "S (replicate n a)"
    proof (induction n)
      case 0
      show ?case using S.Nil by simp
      case (Suc n)
      value ?case
      from `balanced (Suc n) []` have False by simp
      (* thus "S (replicate (Suc n) a)" by simp *)
      (* thus ?case by simp *)
      then show "⋀n. (balanced n [] ⟹ S (replicate n a)) ⟹ balanced (Suc n) [] ⟹ S (replicate (Suc n) a)" by simp

最后一个show之后的命题是从Isabelle / jedit中的证明状态复制而来的。但是,Isabelle报告错误(在最后show):

   Failed to refine any pending goal 
Local statement fails to refine any pending goal
Failed attempt to solve goal by exported rule:
  (balanced 0 []) ⟹
  (balanced ?na3 [] ⟹ S (replicate ?na3 a)) ⟹
  (balanced (Suc ?na3) []) ⟹
  (balanced ?n [] ⟹ S (replicate ?n a)) ⟹
  (balanced (Suc ?n) []) ⟹ S (replicate (Suc ?n) a)

现在注释掉的证明目标导致了同样的错误。如果我更换0Suc的案例,则show案例的最后0出现错误,但Suc案例不再出现错误。

有人可以解释为什么伊莎贝尔会在这里接受这些看似正确的目标吗?我怎样才能以Isabelle接受的方式陈述子目标?是否有一般参考当前的子目标?我认为考虑到我使用的构造,?case应该做那个工作,但显然它没有。

我发现this Stack Overflow问题提到了相同的错误,但问题有所不同(定理存在一个等价,应该通过{{1}的隐式应用将其拆分为方向子目标在我的情况下,应用提供的解决方案会导致错误和无法实现的目标。

1 个答案:

答案 0 :(得分:1)

您只是错过了内部入门证明中的next

lemma
  fixes n xs
  assumes b: "balanced n xs"
  shows "S (replicate n a @ xs)"
proof -
  from b show ?thesis
  proof (induction xs)
    case Nil
    hence "S (replicate n a)"
    proof (induction n)
      case 0
      show ?case using S.Nil by simp
    next (* this next was missing *)
      case (Suc n)
      show ?case sorry
    qed
    show ?case sorry
  next
    case (Cons a xs)
    then show ?case sorry
  qed