I believe that a similar idea is a solution. From my perspective,
the very notion of "cataloging" runs contrary to the nature of information
in the digital world. Information is its point of access. This notion is
cleary understood by the major search engines like yahoo, google, et al.
The information included at the header of any audio file could be all that
there is for the purposes of identification. That information is perhaps a
file unto itself, but it need not be isolated in a separate database.