Representing defining and using ability competency and similar concepts
For brevity, the term "ability" is used in these pages, and intended to cover the related concepts of skill, competence, competency, and other similar concepts. On the grounds that knowledge can be seen as the ability to answer questions, knowledge is also included under ability.
Any definition of an ability must be able to be represented by a URI. It is most useful if any such URI leads to useful information. As in the W3C's "Cool URIs" document, ideally the URI used for the definition should be distinct from, but resolve to, a URL which gives information about the definition. Even with no other provision, this will allow two systems to know if they are talking about exactly the same definition. This may work where only one definition is used for any one ability concept.
A human readable definition returned when resolving the URI should enable a human to know whether the definition is suitable for any particular purpose in mind, including the reuse of the URI in another system. In an ideal world, people would search thoroughly for such definitions, always adopt them if suitable, and never duplicate definitions that are interchangeable. However, this is indeed idealistic.
If people, as expected, do not always search for and reuse existing definitions, different people and authorities will frequently define terms with the same or similar meanings. The natural state of definitions is for them to be splintered and messy, rather than well-ordered and coherent. In addition to the definitions themselves, what is needed is sufficient information on how the definitions relate to each other, so that as far as possible related definitions can be processed together automatically and meaningfully.
Processing and using such definitions in the real, complex world where people have multiple similar definitions requires a degree of machine processability of this information. It will need to be represented in a way that is able to be understood by the interrogating software. The assumption of the present work is that the basic common means of representing this information will be according to the Semantic Web's RDF. Other general purpose knowledge representation languages are also possible, but not covered here.
For RDF to be available to appropriate software, there are two fundamental possibilities.
The second of these possibilities has a number of evident options: see fuller details in the technicalities page
It is relatively well understood that an ontology can be given in RDF, relating a set of definitions to each other. However, the point to be emphasised most here is that the most important information to supply is that relating the definition to other similar definitions. It is in this way that the inherent splintering of definitions can be tied together again.
The suggestion put forward here is that there is a small set of common relationships between definitions from different authorities that would allow sufficient automatic processing to be highly worthwhile. These are the same relationships that occur in many different settings, and are the subject of the provisional SKOS mapping properties. These are stated briefly here: the detail is left to the SKOS section of the Technicalities page.
Work on which these relationships are based is discussed in the related work page.
The other SKOS "semantic relations" (not the mapping properties) can also be used, alongside other RDF and RDFS constructs, for much of the internal structure of a set of definitions.
Next: requirements and philosophy.
|index||web||publications||Simon Grant||© 2009 07|