[TopicMapsInLIS] Inaugurating the list

[Alexander Johannesen]

On Nov 28, 2007 3:50 AM, Liliana Melgar E. <lilimelgar at gmail.com> wrote:
> Do you know something about the AACR3 and the future of MARC?

I know a lot of serious catalogers disagree strongly with the way RDA
is going, and from what I've seen RDA is still going down a path
that's hard to duplicate through automatic means (meaning, lots of
human work to edit and maintain the meta data, lack of data typing,
high-level semantics of data, low-level semantics of fields, etc.). I
know recent flurries of protests have steered them in a slightly
different direction, but I doubt (very much) that it'll be of
significance.

As to MARC, I think we're moving towards a diverse set of standards as
opposed to cramming everything into the one. That's a good thing, of
course, but the focus on MARC is in itself a bit of a red herring;
there's really nothing wrong with MARC per se as a protocol ; the
problem is with how people use it and what they stuff into it. The
rules of AARC2 and RDA (and others) have focused all too much on low
density of semantics in the field structures, giving us a too high
degree of semantic flexibility within the meta data themselves. It
basically means that almost anything within the fields can be
interpreted in some way. An example would be a field that specify a
date. How many ways are there to enter a date? Quite a few, and
although there are guidelines to dates there's no hard rules, and
certainly none that has been enforced. When we say date ranges, it
gets even worse. Character encodings? laughable. We of course here can
blame tools as much as people, of course. In order to translate rule
sets into machine understandable formats (and not just readable like
now) we need tools and technologies that offer us validation and
strong typing, and there is absolutely no such thing in the library
world as of today.

The future will be lots of formats, with lots of different semantics.
This is hopefully ok, but we do need to agree on our ontologies to
make it work.

> It seems to me
> that the efforts are divided in the library community, and most of the time,
> the new initiatives want to invent the wheel again and don't take into
> account previous efforts.

That's a library tradition, so it's not unexpected. :) The library
world were incredibly ahead of time back in the days, back when you
*had* to create your own, and those standards are often quite robust
still today as they were designed. But of course the world has moved
on, and any attempt I see these days coming out of the library world
reinvents some other outside solution (or parts of them), be it MODS,
OpenURL, SRU/SRW (although this one at least reuses established stuff)
and others ; they all taste a bit of half-embracing the underlying
technologies when what is needed is a full embrace and a huge update
of library infrastructure. Of course, this is hard and won't happen
over night, and I'm not even sure it will happen in time as the
library world is hopelessly slow when it's in a position of not
knowing what it should be doing.

> The good thing about TM is that it could bring
> that associativeness... But then, what about RDF?

I won't advocate Topic Maps over RDF. Heck, at this point, *any*
semantically rich metadata standard would be fantastic. MARC still is
what sits in the middle of most stuff, even when there is a
import/export workflow, it all ends up in MARC. it's bizarre.

Librarians knows quite a bit about ontologies and ontology work, and
what tools and what technology they use for that really doesn't
matter. What matters is trying to get them from looking at it to doing
it in any meaningful way, and I haven't seen any moves in that
direction. Nor do I suspect there will be any, either, but that's a
rant for another day. :)


regards,

Alex
-- 
---------------------------------------------------------------------------
 Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps
------------------------------------------ http://shelter.nu/blog/ --------