Previous Next

Problem: Postcoordinated terminologies are complicated

Medicine needs useful terminologies, but terminologies that are simple to use are not useful, while useful terminologies appear to be too complex to be directly useable.

The complexity of useful terminologies arises from four principle sources:

  1. the size of the domain to be covered
  2. artefacts of the technology
  3. trying to satisfy all users
  4. consistent application of semantic choices

Merging all possible users requirements within one product inevitably means that the 2% of content needed by one user sees risks being entirely obscured by the 98% of content included for everybody else.

Semantic choices arise broadly for two reasons: firstly, there is no practical limit to how much detail you can represent in a Terminology , but we don’t have a good methodology for defining in advance how detailed a Terminology must be for a given purpose. How much detail is ultimately included, therefore, is to some extent an arbitrary choice.

Secondly, in a compositional Terminology , there is often more than one recipe (schema) that could be used to construct new, more complex concepts from more elementary ones. Sometimes the best schema can be determined by using proper analysis, for example using philosophical principles; but sometimes the choice must be made arbitrarily.

The set of all schema choices - your ontological commitments - must then be applied consistently across both the Terminology as published and all new ad hoc postcoordinated compositions. Otherwise, computers will not be able to reliably compare different expressions and determine when they mean the same thing or whether one is a kind of another.

Previous Next
Making the impossible very difficult, ©OpenGALEN.org, All rights reserved