Bye Bye HyperText: The End of the World (Wide Web) As We Know It

[Robb Scholten]

I have been noticing this thread in the popular press about the demise of 
indexing as the point of entry for data retrieval.  I agree with Gerry 
that this does deserve some sober reflection.  The question is, can robot 
indexing systems, neural retrieval engines, and relevancy ranked 
full-text retrieval replace the intelligence of a live person scanning a 
chunk of text and assigning index pointers from a controlled thesaurus?

I would love to see someone in ASIS conduct a study where two groups of 
students attempt to answer queries using traditional index searches and 
something like Excalibur or WebWire and compare the results. 

I would not be quite so hasty to declare the concept of hyperlink dead.  
We have yet to develop a retrieval model that can efficiently use both 
the traditional index to locate your first "gem" and then be able to link 
to all other works cited in the first.  This process of data mining 
(whatever name you want to apply to it) mimics rather well how 
researchers locate sources of information using indexes and 
bibliographies in libraries.

On the other hand, perhaps the neural network-type of retrieval will 
actually supercede this centuries-old method of research.  I eagerly look 
forward to giving it a try!

Robb Scholten
Clinical Computing
Beth Israel-Deaconess Medical Center