On 4/23/10, Weinheimer Jim <j.weinheimer_at_aur.edu> wrote:
>
> Alex,
>
> I think this exchange has been extremely valuable since it illuminates an
> area of genuine difference in the views of catalogers vs. programmers.
I agree 100 per cent. This has been the most valuable discussion I have
seen on here for a long time, perhaps ever. Valuable not only in
illuminating the differences in the thinking between catalogers and
programmers, but also in asking challenging questions that finally get down
to the meat of what a catalog really is, or could be (I hesitate to use the
word "should" because that implies there's only one correct way).
I think neither view entirely wrong. Each side is looking at it through
their own lens. There are advantages and disadvantages to a very
semantically structured system as there are advantages and disadvantages to
full-text searching systems. I would say a big disadvantage to full-text is
that search results are noisy. This is because it treats content as dead
data. Matching patterns in strings basically. A semantically structured
system deals in concepts and how they relate to eachother -- not just a pile
of bytes.
What could be helpful is to throw off the MARC blinders and the programming
and computer blinders. Consider what a catalog really is in an abstract
form without MARC or computers or even books. Sort of a general theory.
Can we do that? Has someone ever done that? Perhaps on that level we can
have an improved discussion.
Bernhard Eversberg issued this challenge:
"Please, get real and present workable models. Find outside specialists
who can help with it. Rig up a sizable database to demonstrate the
new concepts and functions. Spread this word, everyone is welcome.
Everyone in this forum will be beside themselves with joy to see it."
I'm seriously taking him up on that. I am working on an idea that I have
been kicking around for many years. It's a generalised semantic mapping
system -- sort of a database of pure relationships between objects. The
objects in this "database" could be anything: Books, patrons, indexes,
concepts, URLs, images, whatever; but the objects reference only indirectly
to what they represent. I am developing this thing without reference to
MARC or even database concepts. I have a crude prototype. I hope that
within a month or so I can present to this group a clearer fleshed out
explanation of what this thing is, what it can do, and perhaps even provide
a working example.
Peter Schlumpf
www.avantilibrarysystems.com
Received on Fri Apr 23 2010 - 09:10:04 EDT