I'd amend it slightly, and I hope Tim would agree: let the humans worry about access, and impact, and making sure appropriate specialists--AI people, librarians, and others-- continue working on getting the computers to do yet better indexing.
On Wed, 12 Mar 2003, Tim Brody wrote: > ----- Original Message ----- > From: "Guy Aron" <[email protected]> > > > ... I would still > > like clarification as to whether the original poster > > was being asked to contribute LC classification > > numbers or LC headings. > > I think the thread started with a concern that choosing LC headings when > submitting an article to an e-print (E-Prints.org?) archive is burdensome. > > From my limited experience catagorisation serves two purposes: > 1) Allowing the location of items within a physical system (e.g. library > shelves) > 2) Locating items on a common theme/subject, as a means of discovery > > 1) is replaced by URIs (URLs, DOIs, etc) in the online age. For the research > literature citations (and citation analysis) should cover 2), supplemented > with free-text searches. > > As the Web has grown it has gone from classification (Yahoo directory), to > boolean-search (Altavista), to graph-based search (Google). Each of these > steps has come about because of a factor increase in the information to > search across, a process that requires decreasing human effort, and as the > classification schemes themselves become so large as to be meaningless. > > I suspect the same will be true of the research literature. > > So the moral of my story is: let the computers worry about indexing, and let > the humans worry about access and impact. > > All the best, > Tim. > Dr. David Goodman Princeton University Library and Palmer School of Library and Information Science, LIU [email protected]
