Thank you in advance
From: dspace-sema...@googlegroups.com [mailto:dspace-sema...@googlegroups.com] On Behalf Of MAlvarez
Sent: Wednesday, July 03, 2013 6:58 PM
To: dspace-sema...@googlegroups.com
Subject: Semantic search technologies and concepts
Hello,
See my replies below:
Hello to all group members, let me first introduce myself: I'm a Java programmer in charge of developing a custom DSpace instance that is considering to use semantic search functionality.
I've seen your addon extremely interesting but before I start appending such code I would like to know first-hand from it's developers which technologies are used:
I would like to know If there is a use of RDF triples to manage such semantic relations, if so :
Item triples do not spawn separate documents. The triples are produced on-the-fly in OWL/XML using an XSLT and based on the OAI qdc output. This the DSpace ontology, containing all triples, and can be accessed at any time using the URL displayed in the Options tab.
Currently they are loaded directly in memory by the specified reasoner. A caching strategy seems to alleviate the “in-memory” effect [1]
In principle yes, exactly because of my first point above that is, triples are generated dynamically from OAI.
Thank you in advance
Best regards,
Dimitrios
[1] Semantic Query Answering in Digital Repositories: Semantic Search v2 for DSpace. Int. Journal of Metadata, Semantics and Ontologies 8 (1), 2013