comp.lang.prolog Google Group

Re: Lambdas in Prolog

2010-03-18T13:41:32Z

To facilitate the inclusion of library(lambda) into Prolog systems
the two-clause BSD license has been added:
[link]

Re: Quintus --> Sicstus

2010-03-18T08:50:42Z

It helps if you post the code fragment and the error message. Someone
is likely to be able to help you with proper advice.
--- Jan

Re: Quintus --> Sicstus

2010-03-18T00:26:24Z

Ulrich Neumerkel schrieb:
Here you go, I guess this is done view list_undefined/0.
From the definition of this predicate I conclude that
it will not detect predicates which will become implicitly
dynamic by means of asserta/1 or assertz/1 during program
execution (as by the ISO standard).
Even if the asserta/1 or assertz/1 does not use a

Re: write a prolog interpreter in prolog (ala Lisp)

2010-03-18T00:14:48Z

I refined it a bit
interpret(','(A,!)):-
interpret(A),!.
interpret(','(!,A)):-
!,interpret(A).
interpret(','(A1,A2)):-
!,(interpret(A1),interpret(A2) ).
interpret(';'(A1,A2)):-
!,(interpret(A1);interpret(A2) ).
interpret('->'(A1,A2)):-
!, interpret(A1) -> interpret(A2).
/*
interpret(X):-

Re: write a prolog interpreter in prolog (ala Lisp)

2010-03-17T23:01:13Z

You have to cut out of a disjunction to make ! work. Otherwise, you
can implement ! via catch/3 and throw/1 - not the most efficient way,
though. The (->)/2 would need special attention too.

Re: Quintus --> Sicstus

2010-03-17T22:59:11Z

In SWI, you load your files. Then you say make/0 to get undefined predicates
listed.

Re: Quintus --> Sicstus

2010-03-17T22:38:22Z

Is there a solution to catch "existence error's" in "compile time"? (I know
this is not precise wording in the case of Prolog)
I have very much old files (from 1982), some of them tries override the
system's predicates, eg. I/O operations, arithmetics. Sometimes the compiler
puts a "permission error", but not allways. And if I delete something

write a prolog interpreter in prolog (ala Lisp)

2010-03-17T21:19:31Z

I tried the following, but it didn't handle cut correctly (I guess,
because it makes separate frame for everything):
interpret(','(A,!)):-
interpret(A),!.
interpret(','(!,A)):-
!,interpret(A).
interpret(','(A1,A2)):-
!,(interpret(A1),interpret(A2) ).
interpret(';'(A1,A2)):-
!,(interpret(A1);interpret(A2) ).

Re: dcg question from Learn Prolog Now

2010-03-17T21:02:03Z

No you don't - the answer is different, but it is printed out the same
because the default settings for printing are such that they LOOK the
same - try w instead of ;
And if your Prolog systems does not produce answers like
X = [the, woman, shoots, the, woman, and, the, woman, shoots, the, woman,
and, the, woman, shoots, the, woman, and, a, woman, shoots, the, man] .

Re: Quintus --> Sicstus

2010-03-17T19:56:26Z

Jan Wielemaker schrieb:
Yes and no.
So I am assuming some GC, so the original list goes
up like a little balloon, to the roof, when it is
not used anymore.
The list contains clauses, and these clauses might
have many uses, not only in choice points, but also
through instantiations when the choice point has

Re: Quintus --> Sicstus

2010-03-17T19:41:19Z

Yes, but sooner or later you need to reclaim the old definition. I'm
not really saying it cannot be done, but it is far from trivial.
No. No sensible system will inline dynamic predicates, but any
implementation should be allowed to inline static code at will.
development problems (re-consult does something like this), but

Re: Prolog Incompletness (Was: Prolog's most illogical feature)

2010-03-17T19:00:38Z

Jan Burse schrieb:
Just had an interesting finding (Michael Brady,
Open Prolog Thesis):
[..]
To enable the cut to handle NR frames efficiently, a
new OPAM register, the NR register, is introduced.
This register always points to the most recently
allocated NR frame. By comparing the cut frame
address with the NR register contents, the presence

Re: Quintus --> Sicstus

2010-03-17T18:09:10Z

Jan Wielemaker schrieb:
Well actually my experience is that the abolish in
a multi-threaded interpreter is not so much a problem
if you DO NOT implement it as some retract all. You
can in one shot throw away all clauses by setting the
clause list of the predicate to empty(*). Can be done
synchronized and should be very efficient, isn't it?

Re: dcg question from Learn Prolog Now

2010-03-17T13:54:41Z

no, the file loaded fine
done.
well, for some reason, when I hit ";" to see the next solution, I get
the same answer repeated 4 more times :
?- s(X,[]), member(and, X).
X = [the, woman, shoots, the, woman, and, the, woman, shoots|...] ;
X = [the, woman, shoots, the, woman, and, the, woman, shoots|...] ;

DES 1.8.1 Released

2010-03-17T12:22:11Z

[Apologies for multiple receptions]
Hello.
In order to cellebrate the resurgence of Datalog in academia and
industry (Datalog 2.0 workhop, just running now), I would like
to contribute with a modest new release of DES fixing some problems.
So, the new release
Datalog Educational System version 1.8.1