File: CASE-SPLIT.html

package info (click to toggle)
acl2 3.1-1
  • links: PTS
  • area: main
  • in suites: etch, etch-m68k
  • size: 36,712 kB
  • ctags: 38,396
  • sloc: lisp: 464,023; makefile: 5,470; sh: 86; csh: 47; cpp: 25; ansic: 22
file content (54 lines) | stat: -rw-r--r-- 2,422 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
<html>
<head><title>CASE-SPLIT.html  --  ACL2 Version 3.1</title></head>
<body text=#000000 bgcolor="#FFFFFF">
<h2>CASE-SPLIT</h2>like force but immediately splits the top-level goal on the hypothesis
<pre>Major Section:  <a href="MISCELLANEOUS.html">MISCELLANEOUS</a>
</pre><p>

When a hypothesis of a conditional rule has the form <code>(case-split hyp)</code>
it is logically equivalent to <code>hyp</code>.  However it affects the
application of the rule generated as follows:  if ACL2
attempts to apply the rule but cannot establish that the required
instance of <code>hyp</code> holds in the current context, it considers the
hypothesis true anyhow, but (assuming all hypotheses are seen to be true and
the rule is applied) creates a subgoal in which that instance of <code>hyp</code> is
assumed false.  (There are exceptions, noted below.)
<p>
For example, given the rule

<pre>
(defthm p1-&gt;p2
  (implies (case-split (p1 x))
           (p2 x)))
</pre>

then an attempt to prove

<pre>
(implies (p3 x) (p2 (car x)))
</pre>

can give rise to a single subgoal:

<pre>
(IMPLIES (AND (NOT (P1 (CAR X))) (P3 X))
         (P2 (CAR X))).
</pre>

Unlike <code><a href="FORCE.html">force</a></code>, <code>case-split</code> does not delay the ``false case'' to
a forcing round but tackles it more or less immediately.<p>

The special ``split'' treatment of <code>case-split</code> can be disabled by
disabling forcing:  see <a href="FORCE.html">force</a> for a discussion of disabling forcing, and
also see <a href="DISABLE-FORCING.html">disable-forcing</a>.  Finally, we should mention that the rewriter is
never willing to split when there is an <code><a href="IF.html">if</a></code> term present in the goal
being simplified.  Since <code><a href="AND.html">and</a></code> terms and <code><a href="OR.html">or</a></code> terms are merely
abbreviations for <code><a href="IF.html">if</a></code> terms, they also prevent splitting.  Note that
<code><a href="IF.html">if</a></code> terms are ultimately eliminated using the ordinary flow of the proof
(but see <a href="SET-CASE-SPLIT-LIMITATIONS.html">set-case-split-limitations</a>), so <code>case-split</code> will ultimately
function as intended.<p>

When in the proof checker, <code>case-split</code> behaves like <code>force</code>.
<br><br><br><a href="acl2-doc.html"><img src="llogo.gif"></a> <a href="acl2-doc-index.html"><img src="index.gif"></a>
</body>
</html>