Key questions on Common Upper Ontology
4/19/2004 8:44 PM
SICoP,
First, thanks for all the invaluable insight.
Here are some reasons why prior efforts to standardize an upper ontology have failed, and why there is hope in the future:
If you are developing a stand-along system, you need a stand-alone ontology, not a domain ontology and not an upper ontology. If you are a developing a range of systems, you need a domain ontology, not an upper ontology. A good upper ontology, with tools, interfaces, etc. will have value, but it isn't really needed. This is the reason there are few ontology developers interested enough in an upper ontology to help develop one, or to even use the ones that now exist.
Here's the hope: The U.S. Army is probably the largest and most diverse organization in the world that is attempting to function as a single enterprise. It has a compelling need for cross-domain interoperability. The Army, from the top, could evaluate and select one of the available upper ontologies and gradually mandate its use. DoD would probably not want to mandate a common ontology, but if they see value in the Army selection, could choose to recommend and eventually mandate its use. The Federal Government also probably will not want to be the body to mandate an upper ontology, since they are so far removed from the systems that would have to use it, but they could follow the DOD lead and gradually recommend, then mandate its use.
If this all occurred, the ontology would be improved, demonstrated, tools made available, experienced vendors available, etc. Perhaps industry would start using the same ontology, and only then could it become an approved standard.
As Chair of IEEE P1600.1 Standard Upper Ontology Working Group, I can tell you our group will not build consensus around a given candidate unless the market first adopts it. But standardization not crucial to the Army, DoD, or the Federal government, since each is a large enough enterprise to gain internal value from a common upper ontology. Of course, the wider the adoption the better, but it is not crucial.
Jim Schoening
US Army Communication Electronics Command DCSC4I