Dear librarian friends

[dsrtao]

Because I have so many of you, and what are you for if not to abuse shamelessly with question about libraries?

When I use my library's online catalog, which I do at a frequency ranging between daily and hourly, I am presented with four search methods: keywords, authors, subjects, and titles. I nearly always use authors or titles, and I am generally happy with the results. I understand keyword searching.

Why does the concept of a subject search still exist? It can't help but present you with useless statistics, such as "there are 14,577 books in the library about the Civil War" followed by the first twenty books written by Arthur Aardvark about that subject. If it led to some sort of taxonomy or shelving system it might at least be excusable for helping the novice librarian point a patron to the right part of the building... but it doesn't.

Shouldn't this be replaced by DMOZ or the original Yahoo! directory or something similar? What am I failing to get?

Comments

MVHO

[msmemory.livejournal.com]

AFAICT, the main use for Subject Headers is for the very determined to find "more like this." So you found this book in the catalog by Smith, on the Civil War, and it has an unusual subject header, I dunno, Civil War (U.S.) - Horses. Now you can look up other books with the matching subject header, including ones that unhelpfully were named "Famous Virginians of the Civil War" but include a chapter on some famous horse.
[example search probably does not exist]

Now, if the cataloging software has made most nouns in the subject headings into keywords, then there's not any point.

Re: MVHO

[dsrtao]

OK, so it's a relic from before modern computing, when indexing was expensive. Thanks.

Re: MVHO

[metahacker.livejournal.com]

hey, if I read comments first, I could skip my morning rant. ;)

Vestigal remains of a bygone era

[metahacker.livejournal.com]

From what I understand library search was invented before usability. (insert usual private rant here) So it's based on "What fields do we have?" rather than "What info are people looking for?" When hashing the table was expensive and time-consuming, writing down human-inferred things like "Subject" was very good; now that generating the table is essentially free and sorting it really is free (in comparison to rearranging 100,000 cards, which I've done...) it makes very little sense.

A related rant is how citations are laid out. Who the hell cares which city the publisher is in in the age of the Internet? Meanwhile, you have to fight through a tangle of info to discover the bits you actually need.

Re: Vestigal remains of a bygone era

[msmemory.livejournal.com]

Oh, I remember taking Cataloging in grad school. Painstaking, fussy, and you always wondered whether the user was going to be able to get what he needed.

Subject in an on line library catalog is almost always a carry-over from the old days of card catalogs.

Now that memory space is darn close to free, you can have as many subjects or keywords per volume as you want. In the Old Days, you were limited - and had to print and file a card for each.

[matildalucet.livejournal.com]

Subject is something I understand. I learned it as a child at the drawers of the card catalog in the school library. I've never entirely figured out how to get keyword to do what I want. Not that I couldn't, I just haven't. I am more tech savvy than some people I know. (And less than a lot of you - I'm starting to wonder if I'm growing moss.) The more a computer thingy looks like something in the familiar physical world, the easier it is to get aged parents to try it for themselves.

[robertdfeinman.livejournal.com]

Libraries all use specialized software and in some cases the company supplying the software also "owns" the database. This is done so that any attempt to switch to another vendor will be expensive and difficult.

Perhaps full text indexing such as is being attempted by Google will make for better searching in the future, but it will take a long time since most libraries have material that will never be included into Google or anyone else's similar effort.

There was a similar issue when card catalogs were eliminated. Some libraries still retain cards for older material because the cost of computerizing them was prohibitive. The rate at which material is consulted drops off rapidly. Things more than a few years old are hardly ever consulted except for classics and those doing scholarly research.

Even if full text indexing will become the norm in the future there still needs to be some breakthroughs on retrieval logic. Google and its competitors still use very elementary systems to determine relevance. It's not clear whether they keep their algorithms secret because of competitive fears or embarrassment.