W3C home > Mailing lists > Public > public-rif-wg@w3.org > April 2008

Re: imports + metadata

From: Jos de Bruijn <debruijn@inf.unibz.it>
Date: Thu, 10 Apr 2008 17:06:22 +0200
Message-ID: <47FE2CEE.9020602@inf.unibz.it>
To: Sandro Hawke <sandro@w3.org>
CC: public-rif-wg@w3.org
Sorry for being late on the imports issue.

Let me first respond to Sandro's ideas:

> 'import' of a RIF document would merge in the rules in that document

No contest here :-)

> 'importMeta' of a RIF document would merge in the metadata and also the
>       triples which encode the syntactic structure (which we haven't
>       standardized but we should, and Axel made a proposal [1]

about importing metadata I don't really have an opinion.  I'm fine with 
it if people think it's useful.
I am a little uneasy about importing the syntactic structure of a rule 
set. I'm afraid it will invite people to work on the (nasty) syntactic 
level, instead of the (nicer) semantic level.

> 'import' of an OWL XML file [2] or an RDF/XML-file which is an
>       owl:Ontology would (conceptually merge in the OWL-DL axioms,
>       ignoring all triples not playing a role in the ontology

As discussed in the last telephone conference, you have a slightly 
different intuition behind the combination than I do.  I would rather 
phrase it in the form of a "data source reference", but would not have a 
big problem with "import".

We do have the issue here of deciding how to treat the RDF/XML file. 
Should we use RDF simple entailment/RDFS/OWL DL/OWL Full?
existence of some owl:Ontology triple is not sufficient, because both 
OWL DL and OWL Full have that.


> 'importMeta' of such a document would give you the triples (ie the
>       triples which encode the syntactic structure of the ontology).
>       I'm not sure how you say you want OWL-Full inference or RDFS
>       inference or something -- I think you "import" rules which
>       implement that inference, but the import is understood to be
>       symbolic -- you're allowed to use your own equivalent reasoner.

Importing the structure of an RDF/XML document seems like an RIF-RDF 
combination using simple entailment. So, I don't really see how this helps.
I could imagine using this keyword to "import" annotation property 
values from an OWL ontology.  When talking about OWL DL, a keyword like 
this could be used to say you want to take the semantics of annotation 
properties into account [1].



Now to come to my proposal for the semantics of RIF imports:

A set of RIF rulesets R is imports-closed if, for any IRI i in an import 
directive in any rule set in R such that i identifies a ruleset r, t(r) 
is included in R.
Here, t(r) is obtained from r by replacing every constant with symbol 
space rif:local  with a new globally unique constant.

The definition of entailment for RIF rulesets extends to sets of RIF 
rulesets in the natural way.



If this semantics is acceptable, we still need to decide whether we use 
the "directives" mechanism, or whether we extend the syntax of RIF with 
the "imports" statement.  I don't have a very strong opinion either way.


Now to come back to the issue of referring to/importing RDF/OWL:

We have two somewhat orthogonal issues:
a- how to refer to the RDF graph/OWL ontology
b- how to decide which entailment regime to use


It would of course be possible to leave both things out of RIF, but I 
think we should at least give people the chance to refer to RDF 
graphs/OWL ontologies.  I also have the feeling that most people in the 
working group would not have a problem with that.
First of all, the shape of such a reference would depend our earlier 
decision between directive/syntax extension.

Then, I see three ways to include such references:
1-  use the RIF "imports" statement.  An advantage would be that we use 
the same key word for all things that are "imported".  A disadvantage 
would be that the term is overloaded.
2- define a new keyword, e.g. "data source reference"
3- define a range of keywords, one for each entailment regime, e.g. 
"importOWLDL", "importOWLFull"

If we would go for option 3, we would immediately solve the second issue 
(b).
Personally, I don't like option 3, because it always requires to specify 
the entailment regime.  I would like to make it optional.
Then, I have a slight preference for option 2, because it matches my 
intuition, but I would not have a problem with option 1.

concerning issue (b):
1- If we want to include references to entailment regimes, we can 
include a keyword "entailmentRegime", and invent IRIs for the respective 
entailment regimes.
2- Another possibility would be to include "switches", i.e., define a 
new keyword for each entailment regime and inclusion of such a keyword 
would mean you want to use it.


Best, Jos


[1] http://www.w3.org/2005/rules/wiki/SWC#Annotation_properties

> 
> 
> I suspect this isn't your style, Jos :-)  -- I don't think it's a 100.0%
> solution, you have a squint a little --  but it has a certain elegance
> which might make it worthwhile.
> 
> Thoughts?
> 
> Oh, I should also note that folks have been working on clearing up OWL
> Imports with a new spec for OWL 1.1 [3].  I haven't read it yet.  (I
> doubt it addresses this perspective -- OWL pretends (sometimes
> awkwardly) to have a subset relationship between dialects, so there is
> less need to care about which one you are using.)
> 
>       -- Sandro
> 
> [1] big table near the bottom of
>     http://www.w3.org/2005/rules/wg/wiki/AbstractModel
> [2] http://www.w3.org/2007/OWL/wiki/XML_Serialization
> [3] http://www.w3.org/2007/OWL/wiki/Imports
> 

-- 
Jos de Bruijn            debruijn@inf.unibz.it
+390471016224         http://www.debruijn.net/
----------------------------------------------
Doubt is not a pleasant condition, but
certainty is absurd.
   - Voltaire


Received on Thursday, 10 April 2008 15:24:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:48 GMT