Comments on: SICP sections 4.4.2 - 4.4.4 http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/ Eli Bendersky's personal website Fri, 21 Nov 2008 21:34:56 +0000 http://wordpress.org/?v=2.6.2 By: kkwweett http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/#comment-116097 kkwweett Sun, 16 Mar 2008 15:03:48 +0000 http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/#comment-116097 In exercise 4.69, the ends-with-grandson rule can also be designed with the reverse rule that was just constructed in exercise 4.68 : (rule (ends-with ?somelist ?somepair) (and (reverse (?car . ?cdr) ?somelist) (same ?car ?sompair))) which hasn't got the drawback of converting somepair to a list. In exercise 4.69, the ends-with-grandson rule can also be designed with the reverse rule that was just constructed in exercise 4.68 :

(rule (ends-with ?somelist ?somepair)
(and (reverse (?car . ?cdr) ?somelist)
(same ?car ?sompair)))

which hasn’t got the drawback of converting somepair to a list.

]]> By: eliben http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/#comment-108784 eliben Sun, 10 Feb 2008 04:01:47 +0000 http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/#comment-108784 Flávio, you're absolutely right. I will incorporate your answer. Thanks. Flávio, you’re absolutely right. I will incorporate your answer. Thanks.

]]> By: Flávio Cruz http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/#comment-108557 Flávio Cruz Sat, 09 Feb 2008 18:43:56 +0000 http://eli.thegreenplace.net/2008/02/09/sicp-sections-442-444/#comment-108557 In exercise 4.71 there are some queries that behave quite differently in those two implementations. For example in the simple-query, if you use these assertions: (dummy a) (dummy b) (dummy c) (rule (dummy ?x) (dummy ?x)) In the "delayed" implementation, the driver loop keeps printing answers (indefinitely), on the other hand without the delay the query system doesn't print anything, because it's trying to satisfy the recursive rule. For disjoint the same observations apply: (rule (dummy-2 ?x) (or (job ?x ?a) (dummy-2 ?x))) In exercise 4.71 there are some queries that behave quite differently in those two implementations.

For example in the simple-query, if you use these assertions:

(dummy a)
(dummy b)
(dummy c)
(rule (dummy ?x)
(dummy ?x))

In the “delayed” implementation, the driver loop keeps printing answers (indefinitely), on the other hand without the delay the query system doesn’t print anything, because it’s trying to satisfy the recursive rule.

For disjoint the same observations apply:

(rule (dummy-2 ?x)
(or (job ?x ?a)
(dummy-2 ?x)))

]]>