Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [Person-ontology] D7 - Which languages are better than OWL?

Title:
    Dear Philippe,

Philippe Martin wrote:
Gian,


  
In reality, you retract then this "argument" at the end of your mail 
but, please, suppress "gpzarri" from the above sentence.
    

I assume that by "retract" you refer to my "specialization" (it is not a
"correction", just a new statement specializing yours). Anyway, I replaced 
"most practical_applications do not require a language as restricted as OWL"
by "most non-trivial practical applications cannot be carried on correctly 
    making use of such a limited language like OWL".

    Thank you


  
- ("in 2007, no W3C_language provides a way to specify in a concept
    definition that some relations are mandatory"
      
 Why in a "concept definition" ? There are many other Knowledge 
Representation structures (e.g., NKRL template) besides the usual (and 
trivial) "concepts". I could suggest then to 
... change "... in a concept definition" into "... in a single definition".
    

I made the two changes relative to "concept definition", although 
"concept definition" seemed a sufficient argument.

   
    Speaking only and always of "concepts" seems too restrictive to me. Anyway, thank you for the change.



  
      argument:
        - ("in 2007, no W3C_language provides a way to specify that some relations are 
            contemporaneous"
              objection: - "RDF permits to use meta-statements"(pm)
                         - "meta-statements permit to specify that some relations are
                            contemporaneous"(pm)
          )(gpZarri (pm))
        - ("a good_general_KRL should provide a way to specify that some relations are
            contemporaneous"
              specialization: "a good_general_KRL should permit to use meta-statements"(pm)
          )(gpZarri);

      
    I find all these sentences at least incomplete. First of all, I 
don't find there my starting point for this discussion: "No W3C language 
provides a way to describe correctly and completely simple 'situations' 
(sorry, John), facts, events etc. that, like "Peter gave a book to 
Mary", are intrinsically n-ary". Secondly if, with your 
"meta-statements" sentences, you refer to proposals like that of making 
use of modal logics (sic!) to try to transform artificially into "n-ary" 
something that has been created as "binary", I don't agree at all. Make 
use of binary concepts to describe static notions (as "mammals" and 
"elephants" and the relationships among them) and use (in general) n-ary 
structures when you must describe dynamic situations, events, actions or 
whatever.
    

Then, please suggest "complete sentences"
- without using the term "situation" given the discussion showed that this
  term is ambiguous and is not really needed; you once used the term of
  "n-ary representations" when faced with objections related to your use
  of the term "situation" but I am not sure this solves the problem;
- without using the term "binary relations" since you admitted yourself that 
  binary relations were not THE problem (and they are used in NKRL),
- preferably without using the term "n-ary" (to avoid ambiguities, I strived 
  to avoid it in my reformulations; these formulations incomplete but 
  they are a beginning).
That way, hopefully, the statements will get to the point, will less lead to 
misinterpretations, and will less give rise to recurring discussions,
which is precise the point of making this semi-formal summary.

    It is difficult to speak about something that is "n-ary" without using the word "n-ary". I could say: "No W3C language provides a way to describe simply, correctly and completely real-world facts, events, actions etc. that, like "Peter gave a book to Mary", have an arity greater than 2". If you don't like "arity", it will then be necessary to return to terms like "semantic predicate", "role" etc.  With respect now to "situation", it was well evident that Barwise & Perry had nothing to do with the discussion in progress, but anyway...



Regarding "meta-statements" (or "contexts", or "boxes"), please read 
John Sowa's last email and then please suggest a better term or expression. 
They are not necessarily linked to modal logics. I do not see their 
use in transforming "artificially into 'n-ary' something that has been 
created as "binary".

    I apologize. I have apparently unduly associated with you a Harold Boley's paper on the Relationships Between Logic Programming and RDF, where he proposes (among other things) something in the modal logic style to upgrade RDF and that was presented at a Conference in Melbourne. 



  
    On the contrary, I can agree with the following paragraphs. Only a 
point: I don't think that "expert systems are better than OWL". I think 
only that, from a "CONCRETE application of the Knowledge Representation 
techniques" point of view, the situation beginning of the '80 is not so 
different from the actual one.
    

I have now removed the argument 
"expert systems of the seventies were more interesting than OWL"(gpZarri)
from the semi-formal summary.

    Thank you once again.

    Best regards,


    G.P. Zarri