NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/comp/lang/prolog/2057

Text File | 1992-11-11 | 1.8 KB | 50 lines

Newsgroups: comp.lang.prolog
Path: sparky!uunet!mcsun!sunic!sics.se!thomas
From: [e-mail] (Thomas Sj|land)
Subject: Re: Difference Lists (was: Rotation, or something...)
In-Reply-To: [e-mail]'s message of 22 Oct 92 09:43:51 GMT
Message-ID: <[e-mail]>
Sender: [e-mail]
Organization: Swedish Institute of Computer Science, Kista
References: <[e-mail]> <[e-mail]>
    <[e-mail]>
Date: Wed, 11 Nov 1992 10:57:10 GMT
Lines: 36


>I don't agree at all that "Prolog supports polymorphism and data
>abstraction better than many languages." In this respect, it's no better
>than Fortran.

Maybe so, maybe not. The confusion surrounding diff-structures often comes
from misunderstandings about the semantics of LP. In order to use such beasts
as diff-lists one should:

1) avoid using unification for anything but argument transfer of terms
   with a "user-defined" "type".
2) define a (possibly unlogical) predicate 'diff_equal(X,Y)' for the
representations used.
3) Keep in mind that Prolog unification knows of no equality theory but the
simplest one, that of the Herbrand universe.

or
4) Possibly use Harald Soendergaards tools, when available.

Prolog gives nice possibilities to define "datatypes", i.e. user
defined representations with their own equality theory, in a much
cleaner way than FORTRAN, where you would have to define a function
procedure like BOOLEAN EQUAL(X,Y) where X and Y should be coded names
for diff-structures (how this is done without pointers and structures
cannot be considered elegant by any standard...).


--
(Alf) Thomas Sjoeland
SICS, PO Box 1263, S-164 28 KISTA, SWEDEN
Tel: [phone]  Fax: [phone]  Tel home: [phone]
Internet: [e-mail] (or the older [e-mail])
......