On Wed, 4 Oct 2000, Karen Coyle wrote:
Transforming the format into something like LDAP is probably pretty
As I understand LDAP (never implemented it but read through it a few
times) it seems the key to making it work is the ability to carve up
directory domains in different ways. Within an institution, this isn't so
hard: you work at whale.edu, in the Medical School, in the Library, you
have a job classification, an email address, you are or are not a member
of the gym. On top of all this, there's a 'common name', right?, which at
many of our institutions is our 'netid' or 'uniqname' or somesuch,
uniquely assigned to any incoming community member. All these
distinguishing administrative 'memberships' are what make searching
efficient as the database grows.
The unique common name problem isn't so hard... IANAC (i am not a
cataloger) but we already use name or name, birth-death for this, right?
We could carry that forward or assign some random integer or whatever.
And I'm not too worried about extra names... cheap disks will not go away
soon, and it was a wise man who said "don't ever throw away data."
But how would we carve up the names at a high level? Is everybody an
author? How far out does a classification scheme with "author, musician,
programmer, artist, politician, etc.", go before it collapses miserably?
Do we identify authors with their nationalities? Birthplace, place of
work, or resting place? With the institutions they might have been
Would any of those distinguishing 'memberships' be useful at all in a
search environment? And how could they be arranged to allow at least a
minimal attempt at not imprinting horrible political decisions on every
new entry? Also LDAP isn't meant to identify items or works, so there
doesn't seem to be a convenient way (and hence no reason) to build such
So this idea in general seems vitally necessary... but cataloging remains
hard. Is there an LDAP expert who can refute/confirm that LDAP and its
ilk are not suited to this task, though they seem the best candidates?
Also is there a cataloger in the house who can point to existing standards
or pieces thereof so we can see what we could lean on in such a project
(maybe some of the detail pieces of the DC efforts?)?
If there were some painless way to insert something into these folks'
cuecat lookup process that said "hey, this is a book, grab an lc record
and slurp the name headings into another db, ldap or otherwise, and
associate the directory unique id with the book record..." ahhh, we'd be
doing everyone a huge service. And this really is our job to do, isn't
it? Who else is trained to do it right, and would be patient enough to
make good tradeoffs?