Wednesday, June 14, 2006
I fell across LibraryThing one evening while surfing for I don't remember what? I was drawn by the exotic scent of organization to put in about 57 of my own books. I have some really nice software I paid for to "catalog" books, and I have tried simple cataloging software that was free. Library thing provides a title search among various countries libraries, as well as searching the AMAZON.COM database, and LibraryThing is web-based, and lives on the web in a Blogger supported site. All you need is the title, author or illustrator name, or ISBN etc. The searching uses Z39.50 protocol across numerous databases (you don't have to know anything about what this means to search, and in fact it should be an invisible aspect of searching) It does require correct spelling - no fuzzy searching here. I paid little attention to the editions I brought in, just the title was the thing because I was trying the "tagging" function. Instead of set Library of Congress subject headings, LibraryThing allows you, the user, to "tag" your individual books with the subjects you feel apply. This is hardly standardized of course, and that's the fun part!
With Library of Congress Cataloging, the cataloger wants to pinpoint the subject of a work so that a heading can be given which describes precisely what the book is "about." Thus a new heading: Distracted Drivers, created when a book only about distracted drivers comes across the desk at LC. So it is not the subject heading which may be searched in a precise hierarchical form, it is the call number assigned the work which is the hierarchical form.
The call number system of the Library of Congress is the most widely affirmed system of categorization of works. There is a book with Dewey system equivalents which point to the LC number equivalent of each Dewey category. Thus, since libraries are arranged by call number, when you go to the library you can browse the shelves in informational clumps. When you search online however, you usually use title, keyword or subject, and the average person has no idea what the LC heading for subjects may be, thus there is a system in place for cross-referencing. If the heading is not an LC heading, you will see a "SEE" reference, and if there are related headings, you will see a "SEE ALSO" reference. You can of course search library catalogs by call number and browse call numbers as well, so you can get a list of works in a given subject area.
I have a theory that today's searchers are putting in a lot of general keyword searches. I know I do, as well as title searches with truncated forms, such as just a few words. The problem here is that on AMAZON you will get about a million hits that you have to wade through. Do people care if they have to wade through a lot of hits? How about if the list was less illustrated and more compact, thus allowing you to roll your eyes speedily down a hundred titles on the page?
Back to the fun of LibraryThing: the tags. These are what they say - metadata tags right in the record with which a searcher will find all the "stuff" with the same tag in the database of LibraryThing. If there is no standard list of tags, pretty soon you are all over the place. Is this good? It well may be good. I found that I gave books tags that perhaps others would not have given those titles, yet when I went back to search, I found that there were others who had had the same general ideas, and thus I saw more titles of works in the same general category of (for example) pulp fiction.
So LibraryThing has energized my thinking and brought a lot of semi-dormant theorizing out of my own wetware database! It feels good to throw off a little dust and rust, and to think about stratified, codified, hierarchified and funified information delivery.