Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site h-aiken.UUCP
Path: utzoo!linus!security!genrad!wjh12!h-aiken!brownell
From: brownell@h-aiken.UUCP (Dave Brownell)
Newsgroups: net.lang
Subject: Re: Self-modifying code (Prolog)
Message-ID: <106@h-aiken.UUCP>
Date: Thu, 12-Jan-84 02:02:55 EST
Article-I.D.: h-aiken.106
Posted: Thu Jan 12 02:02:55 1984
Date-Received: Fri, 13-Jan-84 06:08:14 EST
References: <1096@mit-eddie.UUCP>, <566@sbcs.UUCP>
Organization: U.C. Berkeley
Lines: 18

The reason that the "purists" frown upon such self modifying Prolog
code as counters is actually significant.  One of the great hopes for
logic programming is that it be an efficient road into parallel
processing, such as with dataflow computers.  Self modifying code does
not lend itself to parallel execution; training programmers with
techniques that are already on the way out will make those techniques
calcify like COBOL.  Logic programming should force a different mind
set than serial languages; counters and side effects should NOT be
in the domain of discussion.

Actually, you didn't even pick the most common example of SMC in Prolog:
adding facts to the database!!  Here's an example of a case where to
get useful programs written, you NEED to write "self-modifying" code.
Does any other language make you do this?

	Dave Brownell
	Sequoia Systems (posting from Harvard University, Aiken Lab)
	...wjh12!h-aiken!brownell