Matthew:
Thanks for your thoughts. I understand completely about the
disconnect between what you saw on those materials and how librarians
think about things, and I'm happy to translate. I appreciate it that
you took the time to look at the report and slides, and to respond to
them.
Diane,
Thank you. The ppts and the session notes are really helpful. I
think the work being done on RDA in the DC communities is really
important. You, Robina, Mikael, and many others deserve a lot of
credit for making conversations like these happen. I include you all,
when I say "thank you."
I have three comments to make now, and I think the final paragraph in
the session notes highlights each point.
"Corey Harper emphasized that newer models re-using entity
descriptions via referencing is far more efficient than our former
text-based approach. Robina concurred and felt that another theme of
our consensus would be support of a machine friendly, machine
processable approach. Mikael agreed that machine oriented should be
emphasized, to get away from stenography!"
Yes, I agree that this seems a bit specialized on its face, and I'm
not at all surprised that it was difficult to parse. Interesting to
note, though, that Corey is a librarian (and a good one), so it's not
really a simple "clash of cultures" problem.
First, it is really easy to get confused as people from two different
contexts begin to talk with each other. The first sentence here is
completely mysterious to me, for instance. I have no clear idea of
what could be meant by the contrast between "re-using entity
descriptions via referencing " and " our former text-based approach."
I don't meant to suggest that I think there is not a meaningful
distinction being made here, just that from this wording (and I'm not
blaming the note taker!) I can't tell what it is. This situation is
common when we get out of our familiar conceptual and practical
experiences. I just wanted to remind us all about this aspect of
communicating.
This is a really important issue, and anyone who has spent time as an
authorities librarian (like I have) understands this issue, but
perhaps not as it was expressed above. In libraries, we've become
far too comfortable with the idea that bibliographic descriptions and
authority records (which are really just descriptions of another
entity, if you recall your FRBR) are linked together by the text
string in the bib record. As anyone knows who's had to maintain this
linkage over time, it's not simple to do--because the link text is
also the "preferred label" (in vocabulary-speak) it can change over
time, and it can also be easily entered incorrectly. Some library
systems actually link those authority records to the bibs by
embedding a numeric identifier, which allows the authority record to
undergo any amount of change without necessitating any bib
maintenance, and it's this kind of "non-textual" linking that is our
future. Anyone who's seen those systems understands that the users
still see the text, and so, generally, does the cataloger--the system
makes the substitution for both on display.
I would urge anyone who's interested in these kinds of issues to take
a look at the DC Abstract Model
(http://dublincore.org/documents/abstract-model/) which is, I admit,
tough going in places, but it provides much food for thought as we
contemplate what might work well for us as we relate different
descriptions together, in a way that makes machine manipulation much
easier. I think FRBR says some of the same things, but from a
slightly different viewpoint, and without going as far as the DCAM
does in terms of thinking how these entities might be linked and
packaged.
Second, the rest of the paragraph begins to help me understand what
may be the distinction in the first sentence: the newer models are
better aligned with the needs of machine processing than the older
model, which is aligned with the needs of a human reader navigating
analog information systems, e.g. card catalogs, libraries, book
stores, guides to periodical literature, archives, etc. This is
exactly right, and it is hugely important. One of the things RDA is
aiming to do is make rules that lead to records and catalogs
(entity-descriptions and databases?) that are better suited to
machine processing than our current rules. (So you can see why I
think our conversations with the DC community (and I'm including
IEEE/LOM and similar groups here)
I think we should be careful not to see this as an either/or (as
another poster has mentioned). Making things easier to manipulate by
machine does not necessarily make things harder for humans--and in
fact, may well make them easier, if the applications manipulating and
displaying the information are well designed. And of course, one big
challenge for us in this age of Google is to recall that our users
are not generally other librarians, but information seekers much more
wildly diverse even than librarians! So the notion that we can have
one set of rules to create one resource description that works for
everyone is not really a good place to start. We should be aiming to
provide information that would allow a machine which might know
something about the user or the context to display something more
tuned to the needs of that user or context. The use of roles, for
instance, might help with that.
Your example (in your
presentation) of the citation note is a good one. A note works well
in an analog information system, but in a more automated system it
needs to be more explicitly designated as reference to another
information entity that has a particular type of relationship to the
one being described. Then we can program our machines to act on the
data and provide services for that human reader (or information user)
that assist or enable that person to do whatever it might be the he
or she is trying to do.
Exactly. If you think about why notes are there the way they are, it
has everything to do with the limitations of those 3 x 5 cards we
used to spend our time typing (those of a certain age, that is). The
idea on those cards was very sound at that time--all the most
important stuff would be in the first inch or so, and everything else
relegated to some space further down. This notion of "primary" and
"secondary" runs all through AACR2, and it relegates a lot of our
useful information to places that cannot be parsed by machine, for no
other reason than that we've always done it that way. I don't think
we can afford to keep thinking like that.
Third, re: "stenography!" This one, for me, points to some of the
hardest issues we face in making RDA. What is the right mix of
human-orientation and machine-orientation? Can we provide rules for
both in one set of rules? I think at one level the rules for an
individual describing a resource and relating it to others are the
same as those needed for facilitating a machine-processable approach.
But is a simple, easily used, single presentation of those rules
possible? That is what RDA is trying to be.
To be honest, I think if we're to ensure that we're as important to
the future as we have been to the past we need to stop worrying so
much about balance and start using machine processing in any way that
we possibly can. We need to consider that libraries will be less and
less able or inclined to pay for what we consider "good cataloging"
and that in order to ensure that we get the best information to the
public we need to focus like a laser on how to organize it so that
the machines can assist us and the users in ways we don't even
understand completely yet.
Consider that as the world becomes more digital, full text indexing
will be more available for more materials. It may make more sense
for us to focus more on the kind of metadata that will allow a
machine to make distinctions among various indexed items and help
users to sift through and categorize materials as they search. Now,
we tend to think of ourselves and our metadata as an alternative to
full-text, with that kind of mixed metadata+indexing approach seen as
a threat.
So, by his disparagement of "stenography," Mikael was agreeing with
my points about transcription being something we need to re-examine
as a basic assumption underlying what we do. I can't help thinking
that we'd be better off discussing in more depth some of these larger
issues, rather than going over the specific details with a
fine-toothed comb.
And no, there will not be a single presentation of the rules for all,
at least not the detailed ones. It seems very likely to me that the
various specialized communities will want to continue on after there
is agreement on the general rules, and apply them to their materials
(as they've often done in the past). And, in this context,
traditional libraries ARE a specialized community, not the mainstream
anymore. But it's the agreement on the larger models and general
principles and rules that will bring us the interoperability that we
so desperately need in our future, no matter where we're building our
metadata, and that's where we should be spending more of our effort
now.
Diane
Again, thank you.
Matthew
At 10/23/2006 08:13 PM, Diane I. Hillmann wrote:
Oops, apologies to all, as I neglected to add the attachment. End of
day error, for sure.
Diane
Matthew:
I've attached the report of the meeting, which was very well
attended. The two presentations mentioned in the report are available
as well:
Mine: http://dc2006.ucol.mx/papers/miercoles/16.30/RDA_and_DC.ppt
Mikael's:
http://dc2006.ucol.mx/papers/miercoles/17.0/2006-10-04-Manzanillo-RDA.ppt
I hope if people have comments that they feel free to post them on
the list, and I'd be happy to convey them to DC if that seems useful.
Diane
Diane, please do distribute the document that comes from DC. It
would be great
to see that here.
Matthew
[snip]
--
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
Diane I. Hillmann
Research Librarian
Cornell University Library
Email: [EMAIL PROTECTED]
Voice: (607) 387-9207
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
Matthew Beacom
Metadata Librarian
Yale University Library
130 Wall Street P.O. Box 208240
New Haven, CT 06520-8240
phone: (203) 432-4947
fax: (203) 432 7231
e-mail: [EMAIL PROTECTED]
