Re: Sample Bib-1 record

> Date: Thu, 18 Oct 2001 09:48:47 -0400
> From: "Julie Walko" <jwalko@cqpress.com>
> 
> I'm a consultant working with Congressional Quarterly Press and we'd
> like to be able to make our titles available via Z39.50.  The files
> are in XML with locally created DTD's.  The current plan is to hire
> a programmer to write a parser to convert the XML files to a GRS or
> other acceptable data type.  I've got the Bib-1 syntax, but don't
> know how it's supposed to look in the record.
> 
> Does anyone have sample they'd be willing to share?  Or know where I
> can find one?
> 
> I've got a GRS record from LoC's MARC-SGML program, but I don't
> understand its layout.  It doesn't clearly map to Bib-1.

Julie,

You have a slight problem here: there is not really any such thing as
a "BIB-1" record.  BIB-1 is an attribute set used for searching -- a
vocabulary which allows you to express searches like "I want to know
about books written by Kernighan that have Unix in the title."  Wheras
what you want is a record syntax in which to express your
bibliographic data.

While I am a big, big fan of GRS-1, I have to admit that in the world
that we currently live in, it's not a good choice for transferring
bibliographic data.  The standard for that over the last few decades
has been the various national MARC formats -- USMARC, UKMARC, etc.
There are people out there who can talk to you in the most incredible
detail about MARC records ... and some of them probably will.  You'll
most likely be able to find a clean mapping between your data and the
various MARC fields without too much problem.

A few years ago, we might have hoped that GRS-1, which is altogether a
more expressive format, would take over from the MARC formats, at
least in the context of Z39.50.  However, no-one ever got around to
defining a standard way of representing bibliographic data in GRS-1
records, and of course without a standard way of interpreting it, data
is useless.

In the last few years, XML itself has gained an immense amount of
ground as a record syntax in Z39.50 -- to the extent that its support
is mandated in the Bath Profile, while GRS-1 support is not.  (And
don't let me get _started_ on what a tragedy _that_ is :-)  But this
means that you have another alternative in your quest to make your
data available over Z39.50: you could make XML records available!

There's a catch, of course: the records will not be much use to anyone
if they're expressed in your locally created DTD, so you'll need to
translate them into a standard XML bibliographic DTD.  And the problem
there is (you can see this coming, can't you? :-) that no-one's got
around to defining a standard way of representing bibliographic data
in XML records -- at least, not that I know of.

Your best bet if you're looking to go the XML route might be to look
at the DTD defined in Appendix D of the Bath Profile -- that's
http://www.nlc-bnc.ca/bath/bp-app-d.htm
(eXtensible Markup Language (XML) Document Type Definition for Dublin
Core Simple).  It's not a DTD for bibliographic information _per se_,
but for simple Dublin Core-style metadata.  However, my understanding
of the Bath Profile is that this DTD is intended to express the Dublin
Core "dumbed down" form of bibliographic data, so it's better than
nothing.

In summary: GRS-1 is unfortunately a dead duck; XML might be workable
but it's probably too early to go that route; you really want to
translate your data into USMARC.

Hope that's helpful.

 _/|_	 _______________________________________________________________
/o ) \/  Mike Taylor   <mike@miketaylor.org.uk>   www.miketaylor.org.uk
)_v__/\  "I just died in your arms tonight / It must have been some
	 kind of cheese" -- the Cutting Crew, _Died in Your Arms_

Received on Thursday, 18 October 2001 10:25:28 UTC