Path: utzoo!attcan!uunet!mcvax!unido!ecrcvax!bruno
From: bruno@ecrcvax.UUCP (Bruno Poterie)
Newsgroups: comp.lang.prolog
Subject: Re: !/metacall/metainterpreter
Summary: Cool and the Gang
Message-ID: <623@ecrcvax.UUCP>
Date: 15 Sep 88 14:50:52 GMT
References: <1435@kulcs.kulcs.uucp>
Reply-To: bruno@ecrcvax.UUCP (Bruno Poterie)
Organization: ECRC, Munich 81, West Germany
Lines: 56

In article <1435@kulcs.kulcs.uucp> bimbart@kulcs.UUCP (Bart Demoen) writes:
>This is a reply to article <358@quintus.UUCP>  (R. O'Keefe)
>whose author does not understand why the mentioned metainterpreter does not
>work for programs containing metacall if ! does cut outside the scope of call/1;
>so, I will show it:
[...deleted...]
>only if !/0 does NOT cut outside call/1

First point: i consider that the war O'Keefe/Demoen --> Quintus/BIM is going
	somehow outside of control, and that the metainterpret looks more
	like an innocent hostage, or like a pretext, or both. So why do you
	not choose a place and a time (say, Hyde Park, 1st November, 5 am),
	pick up 4 of us as witness, and try to solve it with some sword?
	I think it is time to stop and cool the thing down, or the net will
	burn soon :-)

Second point: now i am probably so stupid, but i was taught the following
	model for prolog procedures, predefined or not:

		Head<1> :- Goal<1,1>, ...Goal<1,i>, ... Goal<1,n>.
		...
		Head<j> :- Goal<j,1>, ...Goal<j,i>, ... Goal<j,n>.
		...
		Head<m> :- Goal<m,1>, ...Goal<m,i>, ... Goal<m,n>.

	where a cut appearing in the list of goals being the body of the
	clause <j> had as effect to remove all choices created since
	passing the neck of this clause, and including the neck itself,
	i.e. forgetting the remaining clauses [j+1 .. m].

	and that BIPs had the following form:

		BipHead :-
			{ and-written body }

	then the call/1 predicate is defined as:

		call(X) :-
				X.
	
	where X is dynamically replaced by its value. Now considering both
	sub-models, if the value of the variable X is the term ! <atom>,
	then the effect of call(!) is to remove all choices between the
	neck of call/1 and the substitution of X by !, including remaining
	clauses for call/1. But the cut will -and can- NEVER affect choices
	which are olders than the passing of the neck of call/1. If you
	want this can of thing, then you have to use a labelled cut, in
	fact a set of BIPs which as a side-effect update the choice-points
	stack (or whatever you use as a backtrack control structure).
	But then this is completely different from a cut, which is integrated
	into the virtual machine (as well as the if-then-else (A->B;C)).
	Therefore building your critic on the final remark that

		>only if !/0 does NOT cut outside call/1

	is almost surrealistic ;-)