Tom,
You wrote:
"Perhaps the third order way might be to mash things up, pushing updates
to subscribers from a central authority."
My understanding, which may be imperfect, is that the Third Order is
just the opposite: it's total decentralization. Weinberger's book is
called "Everything is miscellaneous." As near as I can understand from
Karen's review, that means information control is thrown out the window
completely. Full texts are apparently thought to replace that.
OCLC Worldcat really already provides a situation where there is a
centralized database of "master records" that can be updated to some
extent by member libraries. They also offer a service that provides
updates whenever a master record you have used is improved.
To tell you the truth, the matter of consistency is one area where I may
differ with Thomas Mann somewhat. I don't think it's absolutely
necessary that books and other resources always be indexed consistently.
When someone published an article ("Cataloging must change!") about 15
years ago, showing some evidence that subject cataloging is
inconsistent, he responded to show there were some serious errors of
fact and analysis in the article. Then he went on to say we need to work
be more consistent. I'd say consistency is a goal, but it's not that
fatal a matter if similar books are not necessarily indexed the same
way.
Certainly, as a cataloger, if I find we have two editions of one title
and that they aren't cataloged with the same subject headings, I make an
attempt to correct that. But my point is--and I think this is generally
consistent with Mann's position--that the presence of controlled subject
headings and subject browse screens is going to help people find what
they want even if there isn't total consistency.
--Ted Gemberling
-----Original Message-----
From: Next generation catalogs for libraries
[mailto:NGC4LIB_at_LISTSERV.ND.EDU] On Behalf Of Tom Keays
Sent: Sunday, May 20, 2007 5:04 PM
To: NGC4LIB_at_LISTSERV.ND.EDU
Subject: Re: [NGC4LIB] "Third Order"--was Libraries & the Web
Caveat lector: I'm a public services librarian not a cataloger, but
I've been thinking about this discussion and thought I'd throw out my
example -- a simple one since it is for a book that has only had one
edition but exists as both print (paperback) and electronic (Safari
e-book). So, other than considerations concerning the format, you
would expect the cataloging details to be pretty straightforward.
Morville, Peter.
Ambient findability /
Sebastopol, Calif. ; Farnham : O'Reilly, 2005.
Library of Congress's online catalog currently has these subject
headings for the paper edition:
Database design.
Database searching.
Information storage and retrieval systems.
MPOW used the the following headings for the paper edition. This set
of headings was the most common one I found when spot-checking (via
WorldCat.org links) how other libraries had cataloged it and I suspect
this is how OCLC's primary record treated it (I haven't looked as it
doesn't actually matter):
Database design.
Database searching.
Information storage and retrieval systems -- Design.
Curiously, for the electronic edition from Safari Books, the record in
MPOW's catalog used this completely different set of headings. This
was very strange to me since, other than format, the print and
electronic editions are identical in content and organization. A kind
person would say this shows the flexibility of what's allowed in
cataloging to tailor things to local conventions and a snarky one
would say it shows the inexactitude and capriciousness of cataloging.
Ultimately, I suspect that these were the out-of-the box headings from
OCLC attached to the electronic edition.
Internet searching.
Information retrieval.
Internet--Social aspects.
Looking further, I saw that other common groupings of assigned
headings in other libraries included these:
Database design.
Database searching.
Information storage and retrieval systems -- Design.
System design
Information retrieval.
Internet searching.
Database design.
Database searching.
Information storage and retrieval systems -- Design.
Database design.
Database searching.
Information storage and retrieval systems -- Design.
Internet searching.
Information retrieval.
Internet--Social aspects.
I'm guessing that most libraries used the record(s) supplied by OCLC
pretty much as is, while a small number noticed that the print and
electronic editions had non-overlapping sets of headings and altered
local records to "improve" them.
So, even with a work that has only one expression and two
manifestations (if I understand the FRBR way of expressing it) the
point can be made that all this local diddling with the record is not
an efficient way to go about things.
What's the situation going to be when FRBR moves into the library
mainstream?
Just from a workload point of view, given the levels of complexity
that FRBR will introduce to track and reconcile additional
manifestations and expressions of works, how can we expect libraries
to do better than they are now? If we can't even deal consistently
with a single manifestation, what changes will we have to make to our
approach to accommodate FRBR?
Perhaps the third order way might be to mash things up, pushing
updates to subscribers from a central authority.
Here's a scenario: you could dynamically pull the cataloging and
headings used for local display and search from a global source record
(from OCLC or wherever). As improvements are made to the global record
(by whatever process, consensus or not), they are reflected
automatically (although not necessarily in real-time) in the local. I
could imagine doing this using an RSS feed where the local
instantiation of a given record subscribes to the corresponding global
record in order to be automatically notified of updates. Updates
themselves could be then automatically loaded into the local system --
like SUSHI for cataloging -- using SRU polled against the global
record or some such. Libraries could lock local records, as needed,
either in whole or in part to prevent updates from overriding specific
local enhancements (this being, hopefully, the exception rather than
the rule in the workflow).
Concern has been expressed here about quality of FAST records. I don't
know enough about them to really know if this makes sense or not, but
if one library contributes an enhancement to a given FAST record and
that change passes some sort of approval process then, under such a
system, every library subscribed to that record would receive the
update.
Or we could embrace the world foretold by WorldCat Local and hope for
the best.
Tom
Received on Sun May 20 2007 - 19:31:32 EDT