LC Responses/FH4

From OWL
Jump to: navigation, search
OWL2 introduces a new treatment anonymous individuals (explicitly naming

them). This change is poorly motivated (if motivated at all), no supporting use-case is listed. It is a matter of taste whether this is an improvement or not. Given this, we suggest to remove this additional departure from OWL1.

An addendum to the original comment says:

The OWL2 doc. on "New Features and rationale" says in section 2.7.3 that
 In contrast, in OWL 2 anonymous individuals are identified using node IDs.
   PropertyAssertion( a:city _:1 a:Quahog )

If I understand it correctly, this notation introduces a name for the "anonymous" individual (namely _:1), so that we can then later refer to this "anonymous" individual.

I'm not in itself objecting to this new syntax but (a) it's not clear from the doc. whether the OWL1 syntax is still allowed, and (b) no reason at all is given why this new syntax is introduced

See also the first, second and third discussion threads on the list.


To: Frank van Harmelen <Frank.van.Harmelen@cs.vu.nl>
CC: public-owl-comments@w3.org
Subject: [LC response] To Frank van Harmelen

Dear Frank,

Thank you for your comment
     <http://lists.w3.org/Archives/Public/public-owl-comments/2009Jan/0037.html>
on the OWL 2 Web Ontology Language last call drafts.

We also note the 'addendum' to your original comment in

<http://lists.w3.org/Archives/Public/public-owl-comments/2009Feb/0014.html>

And we thank you for helping us avoiding further confusion on this issue.

It is important to note that nothing changed on the RDF side, and that the treatment of anonymous individuals in OWL 2 is fully backwards compatible with that in OWL 1. Even on the structural syntax side, there is no change in expressive power, but we restructured the syntax to be in closer correspondence with RDF graphs to make it clearer that anonymous individuals are in direct correspondence with blank nodes. In the example you mentioned, for example, the "_:1" simply represents a blank node in the RDF graph, and this is indeed the only syntactic means to define a blank node.

We agree this isn't made very clear in the documents, and we will try to improve the presentation. For example, we plan to add some explanatory text into the New Features and Rationale document on the change of syntax.

We hope this answers your concerns on this particular issue.

Please acknowledge receipt of this email to <mailto:public-owl-comments@w3.org> (replying to this email should suffice). In your acknowledgment please let us know whether or not you are satisfied with the working group's response to your comment.

Regards,
Ivan Herman
on behalf of the W3C OWL Working Group



CUT AND PASTE THE BODY OF THE MESSAGE (I.E. FROM "Dear" TO "Group") INTO THE BODY OF AN EMAIL MESSAGE. SET THE To:, CC:, AND Subject: LINES ACCORDINGLY.

PLEASE TRY TO REPLY IN A WAY THAT WILL ALLOW THREADING TO WORK APPROPRIATELY, I.E., SO THAT YOUR REPLY CONTINUES THE THREAD STARTED BY THE ORIGINAL COMMENT EMAIL




Previous version of reply:


To: Frank van Harmelen <Frank.van.Harmelen@cs.vu.nl>
CC: public-owl-comments@w3.org
Subject: [LC response] To Frank van Harmelen

Dear Frank,

Thank you for your comment
     <http://lists.w3.org/Archives/Public/public-owl-comments/2009Jan/0037.html>
on the OWL 2 Web Ontology Language last call drafts.

First, we'd like to note that the shift in the specification from using truly blank nodes to using nodeIDs is not a change to the *language* but merely a change in the *presentation* of the language. No new capability, esp. from the RDF perspective, has occurred. In the old specification:

    <http://www.w3.org/TR/owl-semantics/>

the patterns of admissible blank nodes (roughly, tree like patterns that can be mapped to nested someValuesOf) could be specified without recourse to nodeIDs due to the (tree-like) frame structure of Abstract syntax constructions. Unfortunately, this is not possible in the new functional syntax, hence we must make use of a different mechanism to express the same point. Additionally, we believe there is value in having explicit nodeIDs, borrowing from RDF, since it makes it clear that there is an actual, significant syntax item there.

Here is an example of the difficulty:

(Remember, in <http://www.w3.org/TR/owl-semantics/syntax.html#2.2>:

   """The syntax here is set up to somewhat mirror RDF/XML syntax [RDF Syntax] without the use of rdf:nodeID.""")

ABSTRACT SYNTAX:

  Individual(anObjectProperty
                Individual(aDataProperty "foo"))

TURTLE:

  [anObjectProperty 
         [aDataProperty "foo"]]

TURTLE WITH NODEIDS:

   _:x anObjectProperty _:y.
   _:y aDataProperty "foo".

FUNCTIONAL SYNTAX:

  PropertyAssertion(anObjectProperty _:x _:y)
  PropertyAssertion(aDataProperty _:y "Foo")

This is essentially an issue of surface syntax used in the specification. The FS is designed such that axioms in the FS correspond to axioms in first order logic (which makes various specification issues easier). This makes it impossible to use implicit BNodes. Furthermore, given the tradition of using nodeIDs in RDF, we feel we are well within best practice and, in fact, think that this style makes what's going on substantively clearer.

Accordingly, the working group has decided not to make the change you've suggested.

Please acknowledge receipt of this email to <mailto:public-owl-comments@w3.org> (replying to this email should suffice). In your acknowledgment please let us know whether or not you are satisfied with the working group's response to your comment.

Regards,
Bijan Parsia
on behalf of the W3C OWL Working Group



CUT AND PASTE THE BODY OF THE MESSAGE (I.E. FROM "Dear" TO "Group") INTO THE BODY OF AN EMAIL MESSAGE. SET THE To:, CC:, AND Subject: LINES ACCORDINGLY.

PLEASE TRY TO REPLY IN A WAY THAT WILL ALLOW THREADING TO WORK APPROPRIATELY, I.E., SO THAT YOUR REPLY CONTINUES THE THREAD STARTED BY THE ORIGINAL COMMENT EMAIL