[Web4lib] full text versus metadata

[Weibel,Stu]

Karen G. Schneider writes

> I think it's the combination of the two types of search--baggy
> (search-based) and skin-tight (metadata-based)--that is most effective

It strikes me that there is a fair amount of
whistling-past-the-graveyard in this notion.  Do we actually have any
evidence that would lead to these characterizations of "baggy" versus
"skin-tight"?

Last year someone pointed me to the Red Light Green project, suggesting
it was really good and worth a look.  I looked.  I searched.  I repeated
my search on Google.

The first page of my search on RLG resulted in 20 hits, if memory
serves, zero of which were actually relevant to what I was searching for
(information on the construction of Japanese lanterns).  My Google
search resulted in 10 first page hits, 7 of which were apparently
relevant and 4 or 5 of which were actually useful in meeting my
information need.

Now, there are LOTS of possible explanations for this, not all of which
lead us inexorably to conclude that the demise of structured metadata is
imminent.  Having spent the last decade of my professional life flogging
the usefulness of structured metadata, my hopes and expectations should
be obvious.

Might the availability of the two varieties of metadata available for
the same set of resources (Google indexing and bibliographic data)
afford an excellent laboratory for the exploration of the relative
merits of full-text indexing and structured metadata?  

Yikes... we might actually find out!

stu

Stuart Weibel
OCLC Research
http://weibel-lines.typepad.org