Path: utzoo!mnetor!uunet!husc6!sri-unix!quintus!ok
From: ok@quintus.UUCP (Richard A. O'Keefe)
Newsgroups: comp.lang.prolog
Subject: Re: Logic of Database Updates, Destructive assignments
Message-ID: <794@cresswell.quintus.UUCP>
Date: 23 Mar 88 00:49:04 GMT
References: <4381@megaron.arizona.edu> <1897@ubc-cs.UUCP>
Organization: Quintus Computer Systems, Mountain View, CA
Lines: 19
Keywords: Reassignable Variables, Database updates
Summary: please explain further

In article <1897@ubc-cs.UUCP>, voda@ubc-cs (Paul Voda) writes:
{ criticising Prolog }
> Quite often the state is encoded by a list, but even if it is a functor
> or a tree there is a significant amount of copying to be done in order
> to create a new state.

{ praising Trilogy }
> Trilogy is careful to preserve the logic in similar situation by copying out the
> value of y into the variable x.				   ^^^^^^^

When does Trilogy require copying?  Does it use reference counting,
as SETL used to, or is it a compile time analysis?

> Trilogy does not require the concept of worlds to to allow the
> updates to the data bases. The input-output variables are used with the
> advantage here.

But the Bowen approach permits the manipulation of several worlds at once.
(Something which isn't possible in vanilla Prolog either.)