Organizing Web Informat

[Gary Fouty]

For those willing to look a bit into the future, there is an interesting
report resulting from the NSF-funded 
Digital Library Initiative --
http://www.computer.org:80/pubs/computer/dli/r50028/r50028.htm, and also
cited as a news item in 'Science' magazine (v.272, p.1419, June 7, 1996).
This experiment in document indexing involved generation of lists of term
co-occurrence in a database of scientific articles.  These lists are then
used in conjunction with the traditional controlled vocabulary.  The
technique is not yet ready for prime time since it required several hours on
an NCSA supercomputer, but then much new technology starts on expensive 'lab
equipment' before it becomes generally available.  I am very sympathetic to
the argument that existing human indexing provides an added value that
cannot be dispensed with at the present time, but suspect that some sort of
machine-based indexing will be necessary in many situations.  As Clifford
Lynch points out in the 'Science' article, new hardware makes practical some
of the indexing approaches that were largely theoretical 20 or thirty years
ago.  As hardware becomes inexorably cheaper and more powerful, and as new
ideas are generated, it seems likely we will see major progress in this area.


Gary Fouty                       Science/Engineering Library       
108 Walter Library              Univ. Minnesota -- Twin Cities      
117 Pleasant St. S.E. #108             (612)-624-1851
Minneapolis MN 55455                   g-fout at tc.umn.edu