🚀 go-pugleaf

In article <6473@goanna.cs.rmit.oz.au> ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) writes:
.In article <3470@geocub.UUCP>, billaud@geocub.UUCP (Michel BILLAUD)
.QUOTES

.> 5. If b is deterministic and followed by a !, you can switch them:
.>      p :- a,...,b,!,c,...     
.>      p :- a,...,!,b,c,...

.This is false.  Consider
.	p :- fail, !.				p :- !, fail.
.	p :- write(yoo-hoo).			p :- write(yoo-hoo).
.	?- p.					?- p.
.    =>	yoo-hoo				    =>	no
.If b is determinate (not "deterministic") and cannot fail, _then_ you
.can switch them.

- My apologises about inventing new words from time to time.
- I suppose the author proposed 5 as a generalisation of the
  conclusion of 4 with the same premises "in the LAST clause".
  So nobody is wrong.

.> From csa09@seq1.keele.ac.uk Mon May 27 12:04:26 1991
.> From: "P." Singleton <csa09@seq1.keele.ac.uk>
.
.> If a (deterministic) goal doesn't instantiate its arguments 
.> any further (perhaps it has useful side-effects), then you 
.> can replace it by
.> 
.> 	\+ \+ goal

.The DEC-10 Prolog library defines gcc(G) :- \+ \+ X.
.It stands for "garbage collecting call".

.> Nota : \+ = not

.Note:  in DEC-10 Prolog and Quintus Prolog 'not' is _not_ the same
.as "unprovable" \+ .  In both Prologs, 'not' is not an evaluable
.predicate, but is defined in the library, and is defined to report
.an error when it won't be sound.  Stick with \+ for portability.

Ok, as long as I dont have to pronounce its name :-) . 
What is soundness in this context ? Does it mean that
the evaluation of not(p(X,Y,foo)) returns an error
if there is no procedure p/3 , and that of \+ p(X,Y,foo)
simply fails ?

Michel Billaud



-- 
Michel BILLAUD                 :  billaud@geocub.greco-prog.fr
Departement d'Informatique     :  
IUT "A", Universite Bordeaux I :  phone W: 56.84.57.92  // 56.84.69.22
33405 Talence  (FRANCE)        :        H: 56.36.65.90 (answering mach.)