On Tue, Jun 22, 2010 at 8:35 AM, Bernhard Eversberg <ev_at_biblio.tu-bs.de> wrote:
> My main question was the other way: what can MODS do that MARC can't?
> And I provided a partial answer: part<->whole relationships.
Well, there's another perceived benefit, as well. MODS was *designed*
as an XML schema and, as such, is theoretically more optimally
designed for the sorts of things people want to do with XML (XSLT, of
course, but also XQuery and whatnot).
The point this raises, though, which is the fundamental flaw in any
"MARC must die"/"This impostor must prove itself to be superior to
MARC" argument is that these are really just serializations. The
problem is, they are serializations that lack any sort of actual
formally defined data model to serialize, so they conflate the two.
The question of "will MODS get us any farther than MARC" misses the
point when it's really just another way to look at, work with and
share the same the data we're looking at, working with and sharing
with MARC (albeit perhaps a better way if you're trying to do
XML-based things). Just like there is no reason to store MARC-as-MARC
in our local systems (although we do), there is no reason should be
storing MODS, either.
If the data model existed, these serializations would just be
domain-specific views of it, /which is how it's supposed to work/.
-Ross.
Received on Tue Jun 22 2010 - 09:32:55 EDT