Friday, October 29, 2010

That's So 1998

"...cracking our brains to force everything into this 19th century view of the world of information called FRBR is a lot like forcing a square peg into a round hole. It just doesn't fit. So, instead of whittling the peg down, or boring out the hole, we should open our minds to new horizons because it's a big world out there."

James at First Thus posted this little snippet the other day. It made me laugh and laugh, because: TRUE.

And it got me to thinking about something I put in my notes at the Karen Coyle webinar. Namely, that FRBR's structure doesn't jive with how we see the "future" of information relationships. Everyone is talking about how information should be a 'web', with everything having the same importance and availability. FRBR doesn't do that, though. From the models I've seen of FRBR, it seems that the relationships only flow one way. Instead of subjects having potentially the same importance as a book itself, subjects are always subservient to bibliographic data. Afterthoughts, add-ons. I know that it's been said that people are working on this aspect of FRBR, presumably to change it, but to me that seems to be a basic building block of how FRBR treats information.

FRBR *is* kind of 19th century. Or 20th century even. I mean, it's a really recent concept that information can reside outside a physical object. When FRBR was published (and let's keep in mind that FRBR was in the works for several years before 1998), the idea that you would put information on the web and nowhere more "permanent" was still nascent. Why would anyone do such a thing? If you had a big idea, why would you just put it out into the ether and not try to get it published, recorded, or printed? Today, that doesn't sound ludicrous at all. But it was only twelve years ago.

Thursday, October 28, 2010

New Models of Metadata Throwdown

I don't usually start writings with this, but: OMG.

I (virtually) attended the first of three webinars yesterday on RDA and the new directions in metadata creation. It was entitled New Models of Metadata and Karen Coyle was the speaker. She was great, of course. I like her writing; she was also a very good speaker.
Anyways, you were probably there, too. There were 300 participants at this thing. At first, things were pretty quiet. Webex offers a chat window in their interface so you can have side conversations or ask questions or whatever. I left at the stroke of 3, when the thing ended, and people were still going on and on about the logistics of changing over from MARC to whatever we may end up with. It was insane. To me, it seemed like everyone was asking Karen herself to explain how SirsiDynix is going to implement RDA. What a ridiculous and useless thing to ask. How could she possibly have any idea?

Also, best moment, which I will recreate here in a vignette:

Attendee: So, are you saying that MARC is dead??
Karen: Yep.


So, I left the webinar feeling very angsty about my fellow attendees, because I thought we got way off track with asking questions about logistics, when Karen was there to talk about new models of metadata. I really wanted to hear her talk about the web structures and how metadata theory is changing. She did get to talk about that, but I feel like the conversation devolved a little.

Now, with the distance of a day, I get it. People are scared. Change is hard emotionally and intellectually, and we're looking at a real sea change in how we create and manage information. MARC has been the only format going for a long time, and AACRII has been the only standard going for a long time, and Karen said yesterday that we're looking at getting rid of BOTH. For many catalogers, who have never dealt with anything else (like, say, EAD and DACS), they can't really even imagine what a different standard and different format would look like. Some of the attendees were having trouble even dissociating the two, often confusing MARC for a standard. There's a lot of education to be done in order to help everyone move forward.

Wednesday, October 13, 2010

What my reference librarian found this morning

So this morning I come walking into the library, coffee in one hand, prepared to walk right back to my desk and commence the wonderful and magical work of fixing something in the ILS. The reference librarian on duty (and in fact, the head reference librarian) yells at me. Yes, she yelled. Yes, I went over and said "Ma'am, we are in a LIBRARY." We both laugh.
Anyway, she directs my attention to this subject heading, and asks "what the heck is this? It's just a number. And there's a whole bunch of them."

651 $7 $a7.150.$2gtt

Yeah. There it was. The dreaded $2gtt marking on that 651. "Gemeenschappelijke Trefwoordenthe-saurus (GTT)" (Joint Subject Headings Thesaurus). In other words: the Dutch.

I'll admit it: I badmouth the GTT like nobody's business. Mostly because I find it to be horrible. What kind of subject heading is "History"? I know that you catalogers out there are with me. It's ridiculous. This is the Dutch National System, not Jan's World of Paperbacks. Surely they can do better? I mean, this is the country that built the dykes, that produced the Dutch Masters, held England and Spain at bay while they built one of the greatest international corporations in the world. Subject headings should be a breeze, am I right?
Anyway, this was a new one, but I wasn't surprised. A number for a subject heading? Why not? Why not just move on to pictograms, GTT?

But really, I am just poking fun at what is really a system that is not designed to be the LCSH. Yes, I made a joke of it to the reference librarian, but at least now when she sees something like that, she will know from whence it came. If you, gentle reader, are interested in knowing more about the GTT, which is really just an index of very general terms and not like the LCSH in either form or function, there's a good article here.

Wednesday, October 06, 2010

Semantic Web BlahdeBlah

Reading about how libraries need to embrace the Semantic Web. Here's my note in the margins to myself:

"Isn't there a larger problem with metadata for the Semantic Web (or XML, or MARC) that all these tags are still *text*, and therefore subject to misspellings, misunderstandings, and the vagaries of the human mind? The problem that we have with databases and their inability to catch inconsistencies is also a problem for any XML or Semantic Web entity. You can call a cat a cat or a feline or a kat or a kitty, and no one can nay-say your tag. But your tag will not necessarily fall in with other tags which may be the same *thing*, but not the same *word.*"

I'm interested in how we resolve this problem. It's a problem that catalogers have obviously been wrestling with since time out of mind, but now we have an infinitely larger population of data-inputters, with infinitely less training in the importance of standardization and uniformity of metadata. Are we forced to just give up and let things be kind of crappy? I don't know that the answer is just "more programming."
"Wicked people never have time for reading. It's one of the reasons for their wickedness." —Lemony Snicket, The Penultimate Peril.