On Tue, 22 Sep 2009 09:14:05 +1000,... wrote:
>I'm not convinced, nor do I see the relevance. If you ask a computer
>savvy guy "what's the best way to store this complex data", the answer
>would not be MARC with AARC2, nor denormalised records in a database.
>And if you asked a cataloger about the best way to record a complex
>title, it would not be a simple title field with a possible sub-title.
You sum it up pretty well. MARC was originally designed in the 1960s to
print catalog cards. I don't believe the video screen had as yet been
figured out, but output only onto paper and cards. Watch the comedy "Desk
Set" with Hepburn and Tracey to see the basic attitudes from back then. It's
funny, and some of it applies to us today.
But MARC was never rethought in terms of database structure. When cards
stopped being printed, many of the problems of the cards were simply
transferred into the computer. Since MARCXML is just another format of the
original MARC, the problems of cards, *and* of MARC, are transferred into
the new world. There are reasons for this, though.
For example, when you mention normalization of data, my heart just drops.
What an incredible undertaking that would be! This is only one area where
our current methods of manual updating fail completely. To combat this, we
have supposedly followed our "standards" but never enforced them, which
means that while there are many excellent records in there, there are also
lots and lots of lousy records, input by untrained, uncaring people. It
seems to me that normalizing this data must be done using automatic methods.
Then as a manager, I can then say if that stuff can be automatically
normalized, what about the new records?
Because libraries inhabited their own habitats, we have been able to ignore
a lot of this, but we can no longer ignore it today. Other entities such as
Google, but lots of other entities as well, want to use our data, but they
won't spend massive amounts of their time to make our information work in
their databases. To make our records fit into their databases is *our task*
if we want to be included.
And that is part of what I have mentioned about losing control.