OwlEd for MaMaS

An OWL Manager for the Match Maker Service

Valid XHTML 1.0 Transitional Valid CSS! Website statistics
SISINF LAB HOME   •   HOME    •      CONTACTS

 
A bit of history...
Yet another Ontology manager? Answering to this question requires a bit of history. Our research group initiated working on logic-endowed e-commerce systems a few years ago. Such research was basically motivated by "the market", that is we were proposed to build a web-based system for matching demand and supply in a students apartments rental scenario, and we quickly discovered that using database techniques and/or text-based information retrieval we were simply unable to catch requirements in an intelligent way. We then reverted to logic, in particular to Description Logics, to try to overcome some of the limitations we encountered. Initially we used the Classic system and carried out supply/demand matchmaking using subsumption service. We soon discovered that standard inference services, though useful, were not enough to catch appropriately approximated matches, which were the most common. We then started investigating ways to rank, based on semantics, matches, modifying Classic structural algorithms. Then we noticed that useful services would be explaining the reasons for a given ranking. To make a long story short we built an inference engine, which we named MaMaS (MatchMaking Service) providing various non-standard inference services, specifically aimed to semantic matchmaking. MaMaS-tng (the next generation) is now a fully functional reasoning engine in a subset of OWL-DL, which supports standard reasoning services, but also other "non-standard" services, which can be extremely useful in a number of applications where approximate matches are needed. MaMaS-tng is DIG 1.1 compliant, but we needed an application independent manager to simply interface with it. After weighing pros and cons of just coding a plug-in for Protegè or building from scratch a new one, we decided for the hard path. OwlEd has started this way. Obviously, while trying to keep it simple and effective, we "built on giants shoulders", trying to catch "best parts" from other ontology managers and editors and to avoid what - as users - we disliked when interacting with them. We also had another advantage, that is we could build OwlEd natively on the standardized OWL language. Currently OwlEd supports not only MaMaS-tng, but also, when using standard inference services, other reasoners endowed of DIG1.1 interface. OwlEd beta is freely downloadable at http://sisinflab.poliba.it/owled/. MaMaS-tng is reachable as an http service at http://dee227.poliba.it:8080/MAMAS-tng/DIG.
TOP
Copyright (C) 2012 SisInf Research Group All rights reserved.