[PLUG] RE: the plight of Beaverton Library

jmplug at sourtrout.com jmplug at sourtrout.com
Tue Sep 23 14:58:02 UTC 2003


* Kevin Theobald <kevintheobald at vzavenue.net> [2003-09-23 00:25]:
> 
> (with typewriters).  At some point (I think 1973), they subscribed to
> a service where they would send the LC numbers of all the books they
> had recently ordered, and they would gat back standardized cards for
> each book (usually around 5 cards per book, since there were author,
> subject and title cards, plus a master card, for each).  Of course,
> that made life much easier for everyone.
> 
> I'm wondering how this is done with online systems.  Does each OPAC
> vendor have a service to update the databases automatically?

No, or at least none that I know of.

>  Or are there centralized servers, with the OPAC vendors providing import
> tools.

Pretty much.  OCLC is a big one.  (LC is a source also.)  I'm not a
librarian, so I'm not completely familiar with what they actually do.  (A
lot of stuff, beyond the scope of this).  Here's my oversimplified (and
possibly not-quite-accurate) summary.  Basically, member libraries
download bibliographic records, make some adjustments to them to reflect
local organization and usage, then import them into their own system. If
a record doesn't exist for an item, they will just create one, and
send it in.  It is probably stamped for approval at some point, and is then
available to everyone.  You could probably draw some analogies between these
types of services and CDDB type services.

>  And is the info in an open-enough format that an open-source
> version could be made?

Version of what?  Central databases?  Or the data itself?

I can tell you that here they log into OCLC via a terminal program (provided
by OCLC) and work with records within their system.  When the cataloger is
finished with a batch of records, they are downloaded to a local file.  We
send that over to our system and import it.

The data is in MARC format, which is a standard.  I'm iffy on the finer
details, but MARC is a standard for both the organization of the
data (logically) and of the files themselves.  They're raw data files that
one can pick through and read with a text editor, but you wouldn't really
want to work with them that way.  There are lots of tools that can convert
them into plain text and back again, and others that let you edit them
directly.  I believe some of them are open source.

Here's a good place to learn about MARC: http://www.loc.gov/marc/

It is also possible to connect and transfer data around library systems via
a standard called Z39.50.  (Z39.50 client connects to a Z39.50 server and
speaks Z39.50.  Gets data, etc.)

Look into that a little more here: http://www.loc.gov/z3950/agency/

 
> (BTW, relying on outside sources for cataloging information had the
> downside that the "authoritative" sources of information were
> sometimes just plain stupid.

Heh, heh, heh.  The head of technical services (which includes cataloging)
sits right next to me.  I get to hear all about them all the time.  :)


> librarian told me, "that's how the Library of Congress did it."  I
> asked if the LoC hired rejects from the Patent office; she didn't get
> the joke.)

I guarantee that my friend over here would be rolling over that one. :)
It's true.  To try to keep standardized, you have to go more or less with
LOC, or not too far from it. She's always griping about strange desisions
coming out of there.

And yes, (in response to the post over on PLUG-TALK), cataloging really is a
Black Art.  I think that's the best explaination I've seen in a long time.


> I think VNC could be useful.  Though I don't know if it is possible to
> have more than one person logged into NT at a time over VNC.
> Certainly the other way (multiple people on Windows logged into a
> Linux box via VNC) works fine.


It can, but they would be sharing the same session, and fighting over
control of the machine.  But even if that wasn't the case, I'm not sure how
this would work for the public, or what it would be intended to do.
Remember, it isn't incredibly simple, it will not fly. :)






More information about the PLUG mailing list