Avoiding Perspective-Relative Silos

From NCOR Wiki
Jump to navigationJump to search

Avoiding Perspective-Relative Silos

Ontologies are often mischaracterized as taxonomies intended to represent some person’s (or community’s) perspective. However, the creation of perspective-relative taxonomies results in idiosyncratic and incompatible views of reality. It is precisely the perspective-relative approach to creating data models and taxonomies that results in data silos.

Furthermore, these perspectives are not accurate descriptions of reality; a building is not a warehouse, a vehicle is not cargo, and a person is not a passenger. Rather, what is accurate to say is that a building is in the “warehouse role” for some temporal period (just as a person can be in a “passenger role”, and a vehicle in a “cargo role”). From the perspective-neutral point of view, a building is “a facility which is a relatively permanent enclosed construction over a plot of land, having a roof and usually windows and often more than one level, used for any of a wide variety of activities, as living, entertaining, or manufacturing” (Merriam Webster's Online Dictionary) The same building can be in any number of roles: house, community center, business, or even a target.

In regards to perspective-relative taxonomies, Sharon Flank states that, “Taxonomies purport to map the way you think, so you can find what you want. Unfortunately, you don’t think the way I think, and your colleagues won’t necessarily think the way you think either ...” (Sharon Flank "Why Taxonomies are Doomed" http://www.docstoc.com/docs/38833312/Why-Taxonomies-Are-Doomed)

Flank goes on to state that: “If your taxonomy needs to support long-term needs in a business or archive, it cannot afford personal idiosyncrasies. But idiosyncrasies are notoriously difficult to avoid … once you have a taxonomy, you aren’t finished. As the world changes, you need to update it and maintain it. Does anthrax belong under Weapons of Mass Destruction, or does it just stay under Diseases?"

The problems identified by Flank are avoided through adoption of the perspective-neutral approach to ontology. The latter creates the conditions for avoiding those perspective-relative (idiosyncratic) and relatively unstable taxonomies, which are the cause of data silos. As has been shown in other domains—for example in biology—the problems identified by Flank can be greatly diminished (or eliminated altogether) by adherence to best practices for creating an ontology. (B, Smith, et al., “The OBO Foundry: Coordinated Evolution of Ontologies to Support Biomedical Data Integration”, Nature Biotechnology,25 (11), November 2007, 1251 -1255. http://www.nature.com/nbt/journal/v25/n11/pdf/nbt1346.pdf)

So, how would a reality-centric (perspective-neutral) ontology treat Flank's Anthrax example? Properly speaking, Anthrax is not a Weapon of Mass Destruction—if this actually needs to be said. A disease ontology would properly define Anthrax as, “an infectious disease of warm-blooded animals (as cattle and sheep) caused by a spore-forming bacterium (Bacillus anthracis), transmissible to humans especially by the handling of infected products (as wool), and characterized by cutaneous ulcerating nodules or by often fatal lesions in the lungs.” (Merriam Webster's Online Dictionary). This same Anthrax Ontology would include Bacillus anthracis as the cause_of Anthrax. Furthermore, the ontology may also include the Bacillus anthracis bacterium in an enhancement_role used in a weapon of mass destruction.