Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!harvard!caip!nike!styx!lll-crg!lll-lcc!well!ljz
From: ljz@well.UUCP (Lloyd Zusman)
Newsgroups: net.micro.pc
Subject: Re: anyone tried Turbo Prolog?
Message-ID: <1042@well.UUCP>
Date: Wed, 7-May-86 20:06:56 EDT
Article-I.D.: well.1042
Posted: Wed May  7 20:06:56 1986
Date-Received: Sat, 10-May-86 14:09:09 EDT
References: <171@athena.UUCP> <133@umix.UUCP> <12159@ucla-cs.ARPA>
Reply-To: ljz@well.UUCP (Lloyd Zusman)
Organization: Whole Earth Lectronic Link, Sausalito CA
Lines: 42

[ "line eater"?  what's that? ]

With regard to Turbo Prolog ...

I'm a bit disappointed.  It's a compiler ONLY, no interpreting allowed.
You must declare all predicates and domains you define.  There are
standard predicates and standard domains (such as "integer", "symbol",
"real", etc.), though.

You can't assert clauses, only facts.  For example,

	asserta(person(man, me)).

... is legal (if "person(symbol, symbol)" is declared previously), but

	asserta((person(X,me):-male(X,me))).

... is not.

Constructs such as

	[1, [dog, cat], [1, 2, 3], ["string"]]

... aren't legal, because lists must be made of items of the same type
(i.e., integers, reals, symbols, etc.).  Plus, even if the list was
modified as follows,

	[1, [10, 20], [1, 2, 3], [999]]

... it would only be legal if its exact shape was declared previously.

I think these considerations reduce the usefulness of Turbo Prolog quite
a bit.  In its defense, it's a slick development system, with a built-in
editor and debugger.  But it's limited enough that I have to consider
it a "toy" Prolog.

I think Turbo Prolog will do a lot to stimulate sales of other, more
complete competing Prologs.  A person from a company called "Arity", which
makes an expensive but fairly complete Prolog implementation for the PC,
claims that his compnay is overjoyed by Turbo Prolog and forsees increased
sales of his product solely due to people getting exposed to Prolog and
then dissatisfied by Borland's offering.