Re: [OEP] Draft of a note on n-ary relations

Excuse me for entering this discussion at a late point in time.
What I'm mostly doing as an N3 user:
take a prolog-like expression: p(a(x), c(z,y))
then I translate this to N3 as:
[:p [:a :x]; [:c :z,:y]].
In prolog the n-arity is syntactically expressed by the parentheses; in N3
by the clams ([]).
The semantic interpretation (at least by my program RDFEngine but I believe
also by Euler and CWM)) is really the same as in prolog (well if by semantic
interpretation is understood the answers that are obtained given certain
queries). The queries have to be in the same format eg [:p [:a ?x]].
When the blank nodes are given a name this amounts to the same solutions as
proposed earlier in the discussion. This is in fact what I'm doing in my
program: I give each blank node a name with format :T$$$x. In a query
however those blank nodes are considered to be existential variables.
Blank nodes are handy when there are a lot of instances eg
[a :payment; :payment_nr "y"; :value "5"; :to_account "xxx"].
In prolog this would probably be: payment(nr, (value,5), (to_account, xxx))
but I think the N3 version is better because payment is a class and not a
property.
Naudts Guido Informaticus
Tel. 02/542.76.01
Fax: 02/538.01.80
E-mail: guido.naudts@just.fgov.be

Received on Friday, 25 June 2004 10:17:59 UTC