Nice paper looking at some of the issues faced by the National Science Digital Library in the process of incorporating from diverse sources. The problems encountered are typical of those documented elsewhere: missing data, incorrect data, confusing data and insufficient data. One way to address these is to perform 'transforms' from the original source, producing a standardized data set. But another, more effective way, is to think of metadata in a different way. "If a metadata record can be seen as a series of statements about resources, then it should be possible to manage the metadata at the statement or element level, rather than the record level. Aggregating both complete and fragmentary metadata from many sources provides the opportunity to build a more complete profile of a resource." The remainder of the paper describes this process using concrete examples.
Today: 3 Total: 6 [Direct link] [Share]
