> If, OTOH, pre-coordinated headings are considered a thing whose time
> has gone, then cataloging using LCSH Red Books can become a whole lot
> easier. Thomas Mann firmly believes otherwise but he may not have lots
> of followers any more.
The trick is how to make something useful out of what we already have.
LCSH sucks, but anyone who thinks a reasonable alternative will emerge
anytime soon has their head in the clouds. It's like designing a new
car -- if it doesn't run on roads that exist today, it has no chance
of succeeding. Likewise, we have to work with metadata that currently
exists if we don't want to redo everything from scratch.
If we ignore the completeness issues with LCSH (i.e. lack of coverage
for some topics) and the inconsistent hierarchical structure, the real
problem is that the rules for constructing headings are Byzantine.
For example:
> .........That same record I quote above has two headings
> that begin "Venice (Italy)", two that begin with topics and are followed by
> "Italy -- Venice"......
That is because AACR2 prescribes how to construct corporate bodies,
and Venice (Italy) is considered a main heading. In the latter case,
Italy and Venice are considered geographical aspects of something else
rather than headings in their own right. By subdividing this way, all
the stuff on a subject with an Italian component would file together
in a card catalog rather than be scattered throughout many drawers.
AACR2 uses a more direct approach so that you don't have to go through
many jurisdictions or areas to get to the name of a small place.
Most of the rules for precoordination make far less sense than this,
specifying which headings can be divided geographically, which can't,
which subheadings can be applied in which situation, etc. Supposedly
this has to do with filing in card drawers, but in practice, this just
isn't true. I have often wondered if some rules weren't developed by
people who needed psychological help, but that's another story.
Even experienced catalogers can't construct these strings without
referring to a manual (and even then it takes them awhile), so the
usefulness of precoordinated headings as they are is limited. However,
they are structured in such a way that you can break the strings into
component keywords and make reasonable educated guesses at their
function.
Form, temporal, geographic, and certain types of subheadings can
easily be detected and it is reasonable to think that they can be used
in combination with other thesauri or appropriately tuned algorithms
to be useful.
kyle
Received on Wed Jun 25 2008 - 10:44:47 EDT