katyn@wam.umd.edu

Assignment 2 - Storage And Retrieval

In terms of facilitating subject access to Web documents (my interest), this can happen on several different levels. For instance, every concept in a document could be associated with an ontological identifier that describes that concept. This would be a fabulous feature for subject access, but it's more realistic that Web documents will be given subject identifiers at the document level, at least starting out. One of the big questions to consider is which classification to use. Way before the Web, all sorts of organizations were in the practice of creating and maintaining their own classifications. More famous classifications include Library of Congress and Dewey for books, and the taxonomy of living things often referred to in the life sciences (of which there are actually several, with the arrangement of the taxonomy one prefers speaking volumes about that person's philosophical orientation).

On the Web, probably the most well known classification is the Yahoo! taxonomy for classifying Web pages. A close second is the Open Directory Project, which allows volunteer subject experts to find Web sites that pertain to their topic of expertise. In both of these areas, the taxonomy (or those who implement it) has control over the way Web sites are classified. After all, if we left classification in the hands of the webmasters, they would carry on like a pack of hungry coyotes fighting over last week's roadkill. Remember before search engines were any good, and a search for "Grandma's cookie recipe" would bring back ten pages of porn, and a search for "mazda clearance" would bring back ten pages of porn, and so on? Unfortunately, as in our brick & mortar society, our electronic society must impose some controls to prevent those whose self-interest approaches sociopathy from making the concepts of fair game and fair market a laughing stock.