On 30/04/2013 03:14, Daniel CannCasciato wrote:
I recently upgraded a record for a gov pub regarding diabetes and nutrition. The kicker? It’s from the Indian Health Service and written for American Indians and had no subject access for Indians of North America or nutrition. Upon further review, a whole bunch of publications fitting these characteristics lacked subject access for their target readers.
This is a serious concern as we enter into the Semantic Web. Very often there are topic-specific databases where everything in the database may be about a specific racial group, or a specific locality or something similar. So, we can imagine a directory of people who live in a certain town. By definition, anyone in that database is within the town so there is no reason to add the town’s name to the metadata for a specific person. Or perhaps there is an “art” database. At the Food and Agriculture Organization of the UN where I worked, one of the most throw-away descriptors you could add was “Agriculture” because that applied to 95%+ of everything we did. So, searching for that is almost as useful as searching for “the”.
So, I can imagine something similar with records from the Indian Health Service: 95%+ would have the subject “Indians of North America” so they would not add it.
The problem is: once you go into the Semantic Web, the “world view” changes with different languages and usages. Suddenly “agriculture” makes much more sense as FAO records go into the Semantic Web. Also, subjects/descriptors suitable to a specialized database may not work the way they are designed to and may result in confusion or offenses, e.g. dyke may refer to a levee, such as in the story of the little Dutch boy.