I began cataloging in the mid-1970s, at the point where libraries
were transitioning from manual and photographic card production to
MARC. Catalogers are a curious lot and wanted to understand how the
coding worked to produce certain outputs on our printed cards. As a
result, many of us demanded the ability to code our cataloging
according to what was then the USMARC/CANMARC/UKMARC etc. standard of
our country rather than a more natural language input screen. In
hindsight, I still think that this was enormously valuable in changing
the thinking of catalogers from the prose paragraph of the catalog
card to a fielded data way of thinking. Whether the container is
MARC, EAD, ONIX, an internal relational database or other structure
doesn't really matter. What is important is that we saw the
granularity that was possible and how it could be used in products
from printed cards to search facets. When new types of resources
became available (think Web sites, e-journals, even physical media),
we could use what we knew of our existing schemas to see why these
resources wouldn't fit into the structures we already had and could
then expand the format to identify this new data and deliver it where
it needed to go. This ability to control our data structures has, to
my mind at least, been invaluable in moving the profession forward and
in keeping us in touch with the IT folks who are coming to our data
from a different direction.
So, while it may be tempting to develop simplified data entry forms,
I would be cautioned about their use with technical services folks.
You need to keep us engaged in the creation, maintenance, and yes the
structure, of metadata. You only have to use a simple template like
LibraryThing to see how easy it is for users to put descriptive
information in the wrong place and thereby diminish the overall
utility of your data. Don't interpret our "pushback" as a negative.
Instead, think of it as our desire to remain engaged in the process.
> There have been a number of posts concerned about data entry for
> bibliographic data which are clearly motivated by the historic tight
> coupling between exchange formats and data entry editors, including
> Jeffrey's below. My comments are interspersed:
>> On Tue, Jan 15, 2013 at 11:22 AM, Jeffrey Allen Trimble wrote:
>> I think that one of the issues I'M attempting to grappling with is not
>> the underlaying structure of the data (I.E. MARC21, marcxml, rdf,
>> etc.) it is the going to be the interface that the professionals use
>> to create the data and maintain the data.
>> The historical hangup is going to be the fact that Catalogers have
>> been using input screens that represent the MARC record (more or less
>> raw, without the Directory data) and over the decades the editors
>> have become sophisticated enough for the user to gain help to the coding.
>> What will be needed to get catalogers to "buy in" to the shift to a
>> new structure, is an editor which the data can be created, manipulated
>> and maintained without know how to use xml coding (or other
>> "<coding>") knowledge.
>> A reminder that MARC was supposed to be "under the hood" and not for
>> the end user-cataloger. No one developed the interface any further
>> and it has become the "defacto" editor.
> This is a really bad, but it's not the way user interfaces work in
> other domains and it's no reason to stifle progress in bibliographic
> data systems. Software systems which assume this coupling may require
> a little more work to bring into the modern era, but it CAN be done and
> it's not black magic or particularly difficult.
>> This is where you will have major pushback from the Librarians that
>> are in technical services more than anything else. Of course, our ILS
>> vendors need to step up to the plate and show us some next-generation
>> models for the creation/maintenance/manipulation with input from the
>> users who create the data.
North Carolina State University