stanbol-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Olivier Grisel <>
Subject Re: [IKS] Persistence Store, Apache Stanbol Migration
Date Tue, 01 Mar 2011 17:04:23 GMT
2011/3/1 Enrico Daga <>:
> Hi,
> I have compiled and run the code from the lite launcher, but then I do
> not know what to do to check the functionalities :)
> Asked for /presistencestore and got a NOT_FOUND message.
> some comments
> On 1 March 2011 11:49, Fabian Christ <> wrote:
>> Hi,
>> looking at the large interface of
>> org.apache.stanbol.persistencestore.api.IPersistenceStore I get the
>> impression that the persistence store is for storing ontologies only.
> I'm can't say which is the intent, but some ontology editing rest
> service is something missing at the moment in stanbol.
>> On the other hand we have the
>> org.apache.stanbol.enhancer.servicesapi.Store interface to store and
>> retrieve ContentItems.
> The difference, if I understood correctly, is that clerezza graphs are
> edited as ontologies, instead of generic RDF graphs, thus applying the
> OWL model.
>> Are these two things related or is the aim of the persistence store
>> somehow different? Or is the persistencestore something that
>> can/should be used in combination with KReS?
> I think we could relate such feature with kres. The ontology network
> manager has the capability to group ontologies and to serve them as a
> single graph, for example, but has no direct editing facility. The two
> components can interoperate by sharing the clerezza store, so
> ontologies stored in clerezza can be edited by the persistencestore,
> combined by the ontology network manager and then served to the
> reasoners component.
> I see also an overlap between the reasoning capabilities of kres and
> the persistence store, which we must overcome in the future (by taking
> the best of the two implementation in one single component, :) ).

+1 we should drop the name persistencestore as it is confusing
(clerezza's TCManager and the underlying Jena triplestore are already
in charge of persisting any kind of knowledge, both TBox and ABox on
the disk).

Let's create a new toplevel component, e.g. stanbol/ontologymanager
that merges the work of both SRDC and CNR on this component.

I let you guys discuss the public API of such a component.

Also please drop the leading "I" in interface names for the sake of
consistency: no other interface in related apache projects (clerezza,
jena, opennlp...) use such a naming convention.

Olivier -

View raw message