Organizing Web Information

[Marc Salomon]

S. M. S. <ssansenb at delnet.net> wrote:
|Why reinvent the wheel?  Why have dozen's and dozen's of different catalogs
|of Web sites?  Why not just set some new AACR2 standards for cataloging Web
|sites, and then let the big boys handle the work?

More like moving from a pneumatic tire to a radial tire (or even those cool
space-age jet packs we were promised in the sixties) than reinventing the wheel
on which the tire rests.

The OCLC model has worked well as a centralized not-for-profit organization
coordinating original cataloging both internal and external to OCLC.  One
reason for this is that there were economic filters that reduced popular access
to paper publishing.  Even with those filters, there are few, if any, sites
that find it practical to add every single OCLC MARC record for centralized
searching due to the sheer volume.  I don't know if there is a single
publicly-accesible server that allows one to search all OCLC records.

If the volume of net resources is orders of magnitude greater than that of
works bibliographically indexed under the auspices of OCLC, then what didn't
scale for searching (not indexing) a moderately large volume of static paper
resources certainly won't for the set of potentially ephemeral net resources
that can theoretically come much closer to an unbounded set than print works.

If it is impossible to walk the web in real time, it is impossible to catalogue
it in real time.  A solution might involve well-known decentralized
domain-specific indexes that could operate at a manageable level of granularity
as opposed to the monolithic centralized clearinghouse.

-marc

--