Every once in a while I am reminded of the 043 marc field, and fantasize about using it in an interface some day.
It includes coded (controlled) information about the geographical topic of the item cataloged.
It appears in a surprisingly large number of records in many of our corpuses, even though hardly any of our systems do anything at all with it; seems like it could potentially be really useful, yeah?
Hey, the code list for marc geographic codes is actually (in rare form) provided in machine readable XML, even.
But, wow, it looks like the relationships in the marc geographic codes are just as odd as the infamous relationships in LCSH.
<gac>
<uri>info:lc/vocabulary/gacs/ff</uri>
<name authorized="yes">Africa, North</name>
<code>ff</code>
<uf>
<name authorized="yes">Africa, Northwest</name>
<uf>
<name>Northwest Africa</name>
</uf>
</uf>
<uf>
<name authorized="yes">Islamic Empire</name>
</uf>
<uf>
<name authorized="yes">Rome</name>
<uf>
<name>Roman Empire</name>
</uf>
</uf>
<uf>
<name>North Africa</name>
</uf>
</gac>
“UF” is a thesaural abbreviation for “Used for”. Normally it indicates a non-authorized “lead in” term, but here some of them are labelled “authorized”? That’s the first weird thing.
But more importantly, am I misunderstanding things, or did that just tell me that “Roman Empire” and “Africa, North” are synonyms? That doesn’t seem right. The geographic area “Africa, North” may overlap with the geographic area “Roman Empire”, it may even be entirely subsumed by it, but surely they aren’t synonyms.
Follow the chain further, and, if “UF” is a transitive property (which I can’t understand any meaning of “used for” that would not be), we seem to be told that “Rome” is a synonym for “Africa, North.” I’m pretty sure that Rome is a city and doesn’t overlap with “Africa, North” at all.
Apparently “UF” doesn’t mean at all what one would assume it does. The question remains whether UF means anything that’s actually useable at all.
Note: Providing XML is good, but you’ve got to also provide some documentation of what the heck the XML means, whether by an XML schema or even just good narration.
Bibliographic Wilderness 
A lot of this can be cleared up by looking at the Name Sequence HTML or PDF. “Rome” is an authorized (boldface) term, but has no code assigned to it. Instead, it has three assigned codes: ff (Africa, North), e (Europe), and aw (Middle East). “Roman Empire,” on the other hand, is not an authorized term, indicating that we should “USE Rome.”
WARNING: I am not a cataloger, nor am I any kind of MARC expert. What follows is pure speculation and inference at an ungodly hour.
I don’t think GACS codes are meant to stand on their own — they’re meant to serve as an adjunct to the geographic subdivision/subject terms in the 6xx fields. So if you’ve got a record containing
651 $a Rome
that record can also have 043 $a e (or $a aw, or $a ff, or any combination thereof in repeating subfields-a).
If it’s got
651 $a Africa, North
or
651 $a Rome $z Africa, North
it could also have an 043 $a ff
The subject authority record for “Rome” lists its compatible GACS codes in an 043 field. If LCCN Permalinks were available for authority records, you could find it at http://lccn.loc.gov/79039816. But they aren’t, so you won’t. (In case you’re wondering, no, the authority record for “Africa, North” doesn’t have an 043. )
See also the MARC Content Designation Utilization website at http://www.mcdu.unt.edu/ Folks from the University of North Texas did a study of what MARC fields were used and not used. Quite extensive, and there are a WHOLE LOT of fields that never get used very frequently
I always point out 043 and 045 to my students, discuss the difficulties of doing geospatial or chronological search in natural or LCSH language, and then sadly allow as how inconsistent use of 043 and 045 and lack of imagination have made for missed opportunities to do some cool things in interface design.
See? I wouldn’t have needed that big disclaimer up there if I’d taken cataloging with Candy like everyone told me to.
Every time I added an 043 to a catalog record I would get annoyed that nothing much was being done with the information and my time spent looking it up!
At our institution we were told not to bother with 043. We never have had, nor does there seem an interest in having, a library system that utilizes that information. I don’t know if one exists. Like Beth said, you can spend a lot of time looking up your codes, but it’s the geographic info in the 6xx’s that gets searched. Add to that the fact that we seem to be looking at ways to merge MARC with non-MARC metadata. In that kind of merge, information that’s not duplicated across all record sets becomes less useful. I don’t know of any non-MARC metadata schemas we’re using that incorporate the 043 codes.
It is a shame. MARC records look clunky to me by today’s data standards, but the granularity of information they contain is very impressive. It’s only in recent times that we’ve had the computing power to mine data at that level of detail, and yet at the same time we are now devaluing the data.
Yep, just as we’re now getting systems that COULD use the 043 if we wanted them to, without too much trouble… we’re going to abandon it and stop using it. Oh well. Really it’s already too late, way too many years of people spending time encoding it and not using it, the systems that could use it didn’t come in time.