Newsgroups: comp.lang.prolog
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!yeshua.marcam.com!usc!howland.reston.ans.net!swiss.ans.net!newsgate.watson.ibm.com!hawnews.watson.ibm.com!syllog2!ulrich
From: ulrich@watson.ibm.com (Ulrich Neumerkel)
Subject: Re: Occur-check considered harmful?
Sender: news@hawnews.watson.ibm.com (NNTP News Poster)
Message-ID: <7795623758-332414@syllog2.watson.ibm.com>
Date: Wed, 14 Sep 1994 17:06:15 GMT
Lines: 15
Disclaimer: This posting represents the poster's views, not necessarily those of IBM.
References: <PERM.94Sep8192108@groucho.csd.uu.se> <9425506.6106@mulga.cs.mu.OZ.AU> <353a5f$jmg@feynman.ee.latrobe.edu.au>
Nntp-Posting-Host: syllog2.watson.ibm.com
Organization: T.J.Watson Research, Yorktown Hts, NY

In article <353a5f$jmg@feynman.ee.latrobe.edu.au> from [13 Sep 1994 14:34:55 +1000]

 |> In much that same way as numbervars/3 is sometimes used to ensure 
 |> (at least) one of the unificands is ground and hence operate as a "cheap" 
 |> occur-check, can't variables in new frames (a la WAM) be initialized to 
 |> something like a frozen var with the appropriate properties wrt binding
 |> and automatic melting?
Something similar, (delaying the initialization of new variables), was
developed by Joachin Beer some years ago. (See JLP, and a volume in LNCS by
the same author).

--
Ulrich Neumerkel, ulrich@mips.complang.tuwien.ac.at
(914) 784 - 7618

