Researchers in multilingual information retrieval and natural lanuage processing are making progress on algorithms for guessing what a text is about and for tranlating queries between languages; soem day we may have reliable programs for recognizing musical melodies or the content of images. in contrast, researchers in metadata focus more on seeking consensus on the meanings of categories for describing resources; on finding ways to alowo simple schemas to interoperate with compelx ones; and on designing frameworks for managing the messy equivalencies between metadata models in different fields and languages. Do the two perspectives form a continuum? What problems do they solve best? For resoruce discovery, what is the best balance between human and machine?
展开▼