RE: comments on RDF mapping

> >   Anyway, this is a long way to say that I second the idea that we  
> > might want to revisit annotations w/respect to allowing a "minimal"  
> > semantics to them - so it woudln't break DL implementations, but  
> > would allow this feature to be more widely used by tools.
> 
> +1
> 

> As you are agreeing, presumably you have some example tools and uses.
> Could you share them with the WG?

I'm sorry, I realised I should have given a more elaborated explanation 
already when reading the e-mail [1]...

Maybe I do not have exactly the same "conceptualisation" of the "minimal 
semantics" expression used in the original e-mail - however, in the following
I'll try to give initial elaboration of the reasons why I supported this, 
together with a sketch of applications that could make use of this "minimal
semantics" of annotations. Note that these views are very roughly outlined
at this stage and open to any discussion and comments - I'm also really 
open to any remarks pointing me to issues and materials that could help me 
with my concerns, since I'm quite new to this initiative and I'm very well 
aware of that.

First of all, I do not object against the fact that annotations should not 
interfere with the DL-based inference in OWL1.1 - I think this should
be separated, indeed. So, "minimal annotation semantics" != logical 
(model-theoretic) semantics of OWL1.1 recommendation in my view, therefore, 
no need for reasoners to try to cope with this "mess". However, I think it 
could be useful to come with a kind of very general granularisation of 
annotations in OWL1.1 - i.e., to propose and model basic annotation "types",
possibly extensible by users. RDFS constructs like rdfs:label can obviously 
be re-used and sub-typed in this annotation hierarchy proposal. I can imagine 
the following example applications from the top of my head:

i) Annotations can be part of the axioms, according to the document [2], 
Section 3. This would allow to connect for instance versioning-related
information (time-stamp, information on version since which this 
axiom has been introduced into the ontology, etc.) not only to the ontology as 
such, but also to particular axioms. Of course, this could be done even now
using rdfs:label or rdfs:comment, but these properties can be used for many
other purposes, too. So, maybe it could be useful to have a common 
versioning-related annotation types in order to let applications make use
of these features in uniformly defined (i.e., standard) way. I cannot 
really estimate practical impact of this, however, it may not be a complete
non-sense in my opinion...

ii) I have certain informal indications among the members of our institute
that a standardised way of uncertainty representation (in OWL) would be a
useful feature for their applications. According to the list referenced in [3], 
these features are de facto considered to be a part of rather far future 
(OWL 2.0), concerning their full incorporation into OWL (presumably into its
model-theoretic semantics). However, having a special annotation type for 
expressing uncertainty (in a uniform way, not by arbitrary use of the
generic annotation properties, again) would allow for this already now.
The DL reasoners can safely omit such information in annotations, but 
applications that would like to make use of uncertainty measures 
associated with particular axioms could actually do it then, utilising a common
way of uncertainty representation in OWL1.1, using a dedicated type of 
annotations. No matter whether they'd use the uncertainty measures just for 
ranking axioms in user interfaces, or for fuzzy-DL reasoning [4], or for 
whatever other purposes. 

Hope I made myself clear enough - feel free to ask for any further 
clarification if needed...

Cheers,
Vit

[1] http://lists.w3.org/Archives/Public/public-owl-wg/2007Oct/0368.html
[2] http://www.webont.org/owl/1.1/owl_specification.html
[3] http://lists.w3.org/Archives/Public/public-owl-wg/2007Oct/0347.html
[4] http://owled2007.iut-velizy.uvsq.fr/PapersPDF/submission_12.pdf

Received on Tuesday, 30 October 2007 18:10:51 UTC