[deepamehta-devel] Apache Stanbol

[Jörg Richter]

No, the use of Stanbol has not yet been discussed. As far as I understand Stanbol's focus is to automatically enrich existing content with "semantics" by the means of several "enhancement engines". This focus is almost contrary to the goal of DeepaMehta which is to set the *user* in the position to think, to reason, and to create. We see the main challenge not in clever algorithms but in the design of configurable task-oriented user interfaces.

We think automatic reasoning and automatic extraction of semantics is of very limited use for the needs of the knowledge worker. We even think the sole use of terms "reasoning" and "semantics" in conjunction with "automatization" is misleading and ignores their original meaning which belongs to the human realm and is not applicable to dead matter like a computer.

However, everybody can use the DeepaMehta framework to develop a module for every use case you can think of. The DM platform might run a "Semantics Extraction" 3rd-party module which in turn could be build on Apache Stanbol. We think the DM data model is well equipped for every kind of knowledge representation. In fact it is more expressive than e.g. RDF as it incorporates concepts like composition, aggregation, and cardinality.

What we do discussing for the moment is to redesign the DM model with RDF as lowest layer and to express the specialities, e.g. DM's ISO Topic Maps flavor as an RDF vocabulary. This could be done by building e.g. on the Sesame or Jena frameworks.
http://www.openrdf.org/
http://incubator.apache.org/jena/

Thank you for your question and the Stanbol hint!
Welcome to the list!

Cheers,
Jörg


On Mar 13, 2012, at 23:17, Dirk Weissenborn wrote:

> Hey,
> 
> just wanted to ask whether it was already discussed or not to build Deepametha on top of Apache Stanbol as Knowledge Management System. In this way Deepametha could focus on more Detailed problems for knowledge representation.
> 
> http://incubator.apache.org/stanbol/index.html
> 
> cheers
> dirk