Re: Gprolog member/2 error ?!

[Pedro Fialho]

> So now it seems to me the input list to pred1/2 is broken. It should
> contain rel(para,_198661,_198705) instead of
> rel(para,_198617,_198705), and this is really the source of your
> troubles... or did I just get confused?
> 

Indeed, the exemplified input list doesn't contain a valid structure for the 
defined constraint. Not sure if I stated this, but the input list to my pred/1
(the [R-C-P|Ls1] in the source code) contains an element which fulfills the 
constraint (found by manual observation), namely:

...,
[[]-sing-[]-E-[],def-sing-[]-F-[],def-sing-[]-G-[],qual-sing-[]-H-[]]-
['conclusao'(E),'diagnostico'(F),paciente(G),nome(G,'Rita Ambrosio'),
rel(de,E,F),rel(para,F,G)]-
[ser(H,a:E)]
,...

I've rewritten my code in a OO way (for clarity) and I still get only
the first element of my input list as result. Namely, for the source:

filterDRS([], []).
filterDRS([R-C-P|Ls], [R1-C1-P1|Ls1]):-
        (aux1(C,P) -> (C=C1, R=R1, P=P1) ; filterDRS(Ls, Ls1) ).

aux1(C, P):-
        memberchk(rel(_,X,Y), C),
        \+ \+ memberchk(rel(_,Y,_), C), 
        nth(1,P,Z),
        arg(2, Z, _:X).

I get the result:

[[[]-sing-[]-A-[],def-sing-[]-B-[],def-sing-[]-C-[],qual-sing-[]-D-[]]-
['conclusao'(A),'diagnostico'(B),paciente(C),
nome(C,'Rita Ambrosio'),rel(de,A,B),rel(para,A,C)]-
[ser(D,a:A)]|_]

Which clearly doesn't fulfill the aux1/2 constraint, although the [R-C-P|Ls]
contains the correct result (above stated).

Thank you very much for your help on this issue. I'm still confused in why 
the list changed...With your hint I got it done/unchanged but my 
understanding of unification still tells me the source should be alright 
(logically/mathematically speaking, the double negation shouldn't make 
difference).

Best Regards,
Peter