From Classical Metaphysics to Medical Informatics

An item in the Wall Street Journal of 13.11.2001 describes how computers might be used as a means of alerting health officials to problems caused by bioterror-ism. The author notes that biological attacks will likely be marked by statistical spikes in the purchase of over-the-counter remedies for specific common ailments , and the data pertaining to such purchases can be made available immediately via the computerized inventory systems used by drugstores. There is a problem however. For while the technology for running databases has reached an impressive state of maturity, the classification systems upon which this technology is based are the products of myriad ad hoc decisions stretching back to the early days of database design. This means that the data in drugstore computers exists in a variety of different forms, reflecting mutually incompatible ways of partitioning the universe of pharmaceutical products. To resolve these incompatibilities is here a relatively simple matter. In many other such cases, however, the inconsistencies resulting from incompatible classifications are leading to ever more intractable problems wherever attempts are made to integrate data from different sources – problems reminiscent of the old fable of the Tower of Babel. Initially the problem of database integration was resolved in case by case fashion. Pairs of databases were cross-calibrated by hand, rather as if one were translating from French into Hebrew. As the numbers and complexity of database systems increased, however, the idea arose of streamlining these efforts by constructing one single benchmark taxonomy into which all of the various classification systems would need to be translated only once. By serving as a lingua franca for