RE: [VM] Configuration management for RDFS/OWL ontologies

Dear Michael and Miles:


After having reviewed substantial amount of related literature I can tell
you that with regard to versioning there is no layered approach. 

In other words there is no layered approach that versions, firstly in a
basic layer RDF triples dealing with the non-trivial problem of blank nodes
in different sources and then on top of this basic layer versions RDF-S and
OWL ontologies.

Michael Klein in his PhD thesis "Change Management for Distributed
Ontologies" has a nice taxonomy of different sorts of changes that can occur
within ontologies but again as I have indicated above he does not deal with
plain RDF data and moreover from a layered approach.

Semversion http://ontoware.org/projects/semversion/ is an initiative part of
the KnowledgeWeb Workpackage 2.3 Dynamics. We are still in pre-alpha version
but hopefully in September/October we will release our first alpha version.
It is a layered approach to versioning on top of a basic RDF versioning
system dealing with b-nodes by using a simple technique called b-node
enrichment (using inverse functional properties).

A quasi-related proposal is found in "Ontology Middleware: Analysis and
Design" a deliverable from the Onto Knowledge FP6 Project by Atanas
Kiryakov, Kiril Simov and Damyan Ognyanov. This is the most related project
to Semversion. Nevertheless their proposal to versioning is quite limited
(as far as one can read in the document's contents).

In fact the Semversion proposal was submitted and rejected for a Semantic
Web related conference by a previous colleague, before being part of the
project. The funny thing is that one of the comments of the reviewer falsely
indicated that what Semversion was proposing was already stated in the
"Ontology Middleware: Analysis and Design" deliverable which proves how many
"light" and inconsistent reviews one can find and be under attack, which
does not justify the paper being rejected for other provable reasons (such
as lack of a "proof-of-concept" implementation for instance in those
moments). 

Although I am getting out of the point my first impression is that
experienced and knowledgeable people do not pay much attention to reviews
(considering that program committees including experts in a given field
normally review papers and also that suspiciously I might state that
currently paper reviews do not have much impact in ones curricula. Hopefully
this might change in the future) whereas people who are attempting to get a
name in their field do try their best. 

Therefore what I would be willing to see in the future is that reviews are
also reviewed by external reviewers and marked according to different
criteria in order to guarantee a minimal quality assurance process such as
it happens in ACM Computing Reviews and moreover that instead of receiving
light and inconsistent reviews there is always positive feedback from
somebody who took seriously the reviewing process. Also this could serve as
main criteria for creating program committees and external reviewer bodies
or committees according to reviewing records. Evidently to guarantee a
non-biased process both papers and reviews should be blind marked ignoring
provenance (organization) and authors.  


Cheers

Carlos   

   

-----Original Message-----
From: www-rdf-interest-request@w3.org
[mailto:www-rdf-interest-request@w3.org] On Behalf Of Uschold, Michael F
Sent: 25 July 2005 18:20
To: Miles, AJ (Alistair); public-swbp-wg@w3.org
Cc: www-rdf-interest@w3.org
Subject: RE: [VM] Configuration management for RDFS/OWL ontologies


There has been some work on this. To find it on Google, try terms like:
"ontology versioning" and "ontology management".

Natasha Noy has recognized that tracking one version to another can be
done using ontology mapping (which is more typically used for mapping
between different ontologies in the same domain, rather than different
version of the same ontology.)

Other work has been done by Michel Klein. 
The idea is to leverage what is known/done in software versioning and
then seeing what if anything is special for doing versioning/management
of ontologies.

Mike


============================================
Mike Uschold
Tel: 425 865-3605              Fax: 425 865-2965
============================================



>  -----Original Message-----
>  From: Miles, AJ (Alistair) [mailto:A.J.Miles@rl.ac.uk] 
>  Sent: Wednesday, July 06, 2005 6:51 AM
>  To: public-swbp-wg@w3.org
>  Cc: www-rdf-interest@w3.org
>  Subject: [VM] Configuration management for RDFS/OWL ontologies
>  
>  
>  
>  Hi,
>  
>  After the discussion on the SWBP-WG VM telecon yesterday I 
>  put down some thoughts on how configuration management for 
>  RDFS/OWL ontologies ought to be done, see:
>  
>  http://esw.w3.org/topic/ConfigurationManagement
>  
>  Has anyone written anything like this down already?
>  
>  Cheers,
>  
>  Al.
>  
>  ---
>  Alistair Miles
>  Research Associate
>  CCLRC - Rutherford Appleton Laboratory
>  Building R1 Room 1.60
>  Fermi Avenue
>  Chilton
>  Didcot
>  Oxfordshire OX11 0QX
>  United Kingdom
>  Email:        a.j.miles@rl.ac.uk
>  Tel: +44 (0)1235 445440
>  
>  
>  

Received on Monday, 25 July 2005 18:36:48 UTC