Re: [DIG-users] Re: Towards DIG 2.0 - suggestions

[A.-Y. T. <tu...@tc...>]

Hello,

I won't attend the workshop in Galway, thus I would like to comment and 
contribute to the current discussion regarding the next version of the 
DIG interface. The idea of dividing the standard into a core and an 
optional part seems to be the most reasonable approach to me for the 
upcoming extensions. Besides the extension to new kind of inferences it 
should also be considered to offer support for standard reasoners for 
"non-standard DLs", such as CEL. It offers classical DL TBox inferences 
for EL with some operators for roles, some of which are not captured in 
the tell statements in the DIG interface.

Regarding the issues mentioned in the paper by Ian Dickinson (that Ralf 
had pointed to in an earlier mail), I would like to stress some of these 
points that we also found problematic when we develop(ed) our 
non-standard inference system SONIC:

a) querying property axioms
The DIG interface should offer functions that determine whether a role 
is transitive or an attribute for given a property name.

b) Concept sets
The concept sets in the response are indeed often clumsy to handle. 
Nevertheless, the information about all synonyms are sometimes important 
for our applications and the concepts sets are a very compact way of 
obtaining this information. I would like to propose to introduce a 
parameter in each query that returns concept names for turning on /off 
the use of concept sets.


To extend the implementation of our non-standard inference reasoner 
SONIC to DIG, it would be necessary to add two new functions to the DIG 
interface.

1) Our application needs to retrieve a concept definition from the DIG 
reasoner. Currently only Racer(Pro) supports this type of queries where 
for a given concept name the concept definition is returned. Of course 
most DL reasoners store the concept definitions in a preprocessed way. 
We need to retrieve these (preprocessed) definitions in our application 
and would like to add such a function to the DIG interface.

2) For performance reasons we need a function to determine whether the 
ontology has changed between two requests. This could be easily realized 
by either a time stamp associated with each classification or a counter 
for classifications since loading the ontology. The DIG interface should 
offer a function for retrieving this time stamp /counter.

Of course, we would like to offer SONIC's inferences (LCS, 
approximation, ...) also via the DIG interface.


Regards, Anni

-- 
*   Anni-Yasmin Turhan, Theoretical Computer Science, TU Dresden   *
*  Phone:++49 351 463 39 167  mailto:tu...@tc...  *


Glen Dobson wrote:
> Hi,
> 
> I have a simple suggestion which I personally would like to see in DIG 
> 2.0 - how about been able to combine boolean queries (satisfiable, 
> subsumes, disjoint) using AND, OR (maybe NOT)? In actual fact my 
> personal requirement is to combine subsumes queries using AND, OR. It 
> would be nice if I could wrap these up in a single Asks document rather 
> than persisting these using yet another format on top of DIG.
> 
> Glen.
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by the JBoss Inc.
> Get Certified Today * Register for a JBoss Training Course
> Free Certification Exam for All Training Attendees Through End of 2005
> Visit http://www.jboss.com/services/certification for more information
> _______________________________________________
> DIG-users mailing list
> DIG...@li...
> https://lists.sourceforge.net/lists/listinfo/dig-users
>