Re: problems with not and unasserted facts

Hello Vic.

On Aug 31, 2005, at 7:12 AM, Vic Bancroft wrote:

> Lindsey Spratt wrote:
>
>> not/1 is not a built-in in gprolog. Instead, use \+ / 1.
>
> This seems to be the most common issue introduced by the ISO  
> definitions !

The not/1 change seems to be a problem for people working from (poor)  
textbooks. I guess a lot of textbooks must use not/1, although I think  
the standards committee probably did the right thing by settling on  
\+/1 instead. Perhaps gprolog (and others) could have some advisory  
when not/1 is used: "use \+/1 instead, it's logically pure, or turn off  
this advisory by setting the i_like_not flag to true."

[I think the biggest failure of the ISO definitions was not including  
term_expansion and DCG. I consider these core features of a usable  
Prolog system.]

>> You could define a not/1 predicate as:
>>
>> not(Goal) :-
>>     \+ call(Goal).
>>
> Hey Lindsey, don't you have a compatibility library of predicates  
> around somewhere ?

mp32compatibility.pl is part of the XGP source:
http://cvs.sourceforge.net/viewcvs.py/xgp/xgp/src/bridge/ 
mp32compatibility.pl
http://cvs.sourceforge.net/viewcvs.py/xgp/xgp/src/bridge/ 
mp32compatibility_ops.pl

It does not include a not/1 predicate though, since MP32 didn't use  
that.

Cheers,
Lindsey