> It's the *potential* that matters and the *intentions* to be served.
> You are right that more granularity is not automatically beneficial,
> but there are always some chances that it turns out that way.
>
> If, OTOH, you could prove that MARC's additional granularity is
> not really and honestly of use anywhere, then by all means let's
> go ahead and finish MARC off.
Anything that exists probably already has _some_ value to _someone_ by virtue of the fact that they felt the need to create it in the first place. What I agree is disingenuous is requiring proof that something is not valuable. I could simply say "prove to me that MODS isn't valuable" and we'd have a nice little self-referential argument. Pointless.
> My main question was the other way: what can MODS do that MARC can't?
I agree this is the better question: what value do these new things bring *compared to the value that is lost in moving to them?*
MARC is primarily a markup format, and MODS is a large (though not flawless) step towards a data storage format. Put differently, MARC does many things, one of which is hold data. MODS focuses on holding data rather than doing other things.
Example: for most top-level elements in MODS, order doesn't matter. Contrast this to, eg. subfields in MARC 245 ($abnp vs $ahbc, etc.) in which the semantics are encoded in a combination of subfield order and ISBD punctuation. A set of <titleInfo> elements *might* appear in any order (and anywhere within the record), but still preserves _most_ of the semantics using XML syntax.
MARC proponents will argue that there are well-defined rules for decoding the syntax using punctuation and subfield ordering, but in 2010, an XML schema is both far more accessible to more technology (both within and outside of libraries) as well as not bound up in a dusty AACR2 tome on the shelf.
Just to throw some meat to the dogs, I would argue that we should consider cataloguing in MODS (and using crosswalks to MARC when necessary for interoperability) as a step forward, and manage the exceptions. That is, assuming we could get a reasonably usable cataloguing system which could do so.
MJ
Received on Tue Jun 22 2010 - 09:33:37 EDT