Mind to Market

Wednesday, April 21, 2010

The Gist of It

The topic of upper ontologies was taken up by Dave McComb in a recent talk at Boulder’s Semantic Web Meetup. Dave is the president of Semantic Arts and the chairman of SemTech; the Semantic Technology Conference. Upper ontologies are ontologies that contain general concepts that can be used across all domains, concepts such as time, space, geographic location, organization, etc. These ontologies are useful in that they reduce the time needed to recreate these general concepts when building domain ontologies and they provide a common set of concepts that can be shared between domain ontologies. If two domain ontologies are developed using the same upper ontology, they will be that much easier to integrate since they will share the same higher level concepts.

Although this may sound reasonable, the world of upper ontologies is far from straightforward. Defining abstract, general concepts to be used across all domains is akin to picking a universal political party or religion, i.e. one size does not fit all. As a result the field is crowded with the likes of Cyc, Basic Formal Ontology (BFO), DOLCE, Sumo, Dublin core, etc. Dave analyzed the ontologies out there and decided to come up with his own: Gist, the minimalist upper ontology.

Dave has attempted to fulfill two main requirements with Gist: cover a very broad range of future applications and cover them with the fewest number of concepts. Sounds good, after all one barrier to using some upper ontologies is the steep learning curve; the adoption of many higher level concepts before even beginning to sort out the domain concepts. Dave’s business centric viewpoint has limited the use of more philosophical concepts and has included the use of business concepts such as property, money, and corporation.

Is Gist “better” than the other upper ontologies out there? That may be a question that is impossible to answer. First, people must agree on what they are looking for in upper ontologies, no small feat in itself, then a tool must be available that would be able to compare the ability of these ontologies to fulfill those requirements, again, not something that is currently available. For this reason, most upper ontologies are chosen based on a recommendation or familiarity, not the most objective reasons for choosing a technical foundation.

Dave is in the ontology building business, which means that he is a power user of upper ontologies. He designed Gist using his own experience and to aid his company in building domain ontologies for clients. Therefore, Gist will either meet his needs, or he will change it or abandon it in favor of some other upper ontology. He is now distributing Gist freely, not quite an open-source project but it may happen.

Labels: , , , ,


Post a Comment

Subscribe to Post Comments [Atom]


Create a Link

<< Home