From whitehal@aosun.cs.uiuc.edu Sat Jan 13 19:33:18 1990
Received: from cs.wisc.edu by steves.cs.wisc.edu; Sat, 13 Jan 90 19:33:15 -0600
Received: from gaea.cs.uiuc.edu by cs.wisc.edu; Sat, 13 Jan 90 19:33:09 -0600
Received: from aosun.cs.uiuc.edu by gaea.cs.uiuc.edu with SMTP
	(5.61+/IDA-1.2.8) id AA11561; Sat, 13 Jan 90 19:32:59 -0600
Received: by aosun.cs.uiuc.edu.cs.uiuc.edu (3.2/9.6),
	id AA03779; Sat, 13 Jan 90 19:32:55 CST
Date: Sat, 13 Jan 90 19:32:55 CST
From: whitehal@aosun.cs.uiuc.edu (Bradley Lane Whitehall)
Message-Id: <9001140132.AA03779@aosun.cs.uiuc.edu.cs.uiuc.edu>
To: shavlik
Subject: last deduce fix
Status: RO

Jude-

There is the final (:-) fix for the DEDUCE code. 
This seems pretty stablenow.  I have a 335 rule expert
system running on it -- lots of cuts and it seems to be
fine.  Before in this routine I was always returning
nil when not a cut, instead of what the called routine
returned (temp).  Enjoy.

Brad
-------------------
(defun gnull (x &aux temp)
  "test when the generator is empty - null list"
  (if (catcher-p (first x))
      (progn
      (setf temp (catch (catcher-label (first x))
		   (null (normalize (rest x)))))
      ;(format t "~%**gnull catch ~a" temp)
      (if (eq 'thrown temp)
	  t
	  temp)) ;want to return  what the user sent
    (null (normalize x))))