Saturday, June 20, 2009

A couple of weeks ago, I discovered LibraryThing. Now I'm sure most of you have already heard about LibraryThing - it seems to have been written up a quite a few places - but somehow I'd missed out. If you haven't heard of it, LibraryThing is an online service for organising your books. Because I have a lot of books, many rather indifferently organised, this was immediately appealing to me.

I chanced to be Googling about, looking for the name of a bookshop in Brisbane, and Google sent me to a page where a group of Australians were comparing notes on Australian bookshops. They mentioned the name of the one I was looking for (which has slipped my memory, because I haven't had to go to Brisbane since 2005,) so I made a note of it and then decided to check what site I had landed on. Behold, LibraryThing: the page I had been reading was a thread in their Talk area.

So I had a poke about, and registered a free account to try it out. LibraryThing will let you load 200 book records for free, so I extracted a chunk of my nonfiction list, re-organised it to suit LibraryThing's bulk upload system (which is pretty simple), and waited while the file got digested by the loader. It actually created 217 records before it quit, and the quality of the records retrieved was pretty good. LibraryThing searches library and online booksellers all over the world - you can configure which sources you want searched - and retrieves records which you can edit to suit your needs and add to you collection.

However, the stuff in my existing nonfiction records was mostly quite recent (within the last five years). I decided to test a bit more rigorously. So I registered another free account, for testing, and went upstairs to select some books. I settled on Nancy Mitford's "Wigs on the Green", which has been out of print for decades; a four-volume edition of Alexander Pope's "Poetical Works", dated 1853; and a privately printed edition of the Heptameron of Marguerite of Navarre, also 19th century. LibraryThing found them all (OK, we had to dig a bit for the correct edition of the Pope, but we got it).

Impressed, I registered a paid account (unlimited books), and ordered one of their "CueCat" barcode scanners. Then I went back upstairs to the room which is labelled "Bedroom 4" on my house plans, but which is, in fact, full of proper steel library shelving, most of it double stacked with books. I got an armful of old Penguins and Everyman editions from the middle aisle, and started loading books into LibraryThing: most of the old books I own don't have a barcode or an ISBN number, so searching for author, title and edition is going to be the only way to load them. Later I uploaded my entire author/title fiction list.

Some observations: LibraryThing works best if you tag your books, at least roughly, at the time of loading. At least tag for fiction or nonfiction. If you are doing bulk uploads, do not chuck a couple of thousand records on the load queue in one lump. First, you will probably cause the queue to hang. Second, you will have no way to easily check what loaded. Split your load into batches of about 50, and tag them by batch number. Then you can easily call up a batch by tag name, and compare it to the file of data you uploaded. This makes it much easier to identify records that don't get a hit, or records that retrieve bad data. I got a couple of very strange books added to my LibraryThing collection, caused by corrupted ISBN numbers in the load file, and it took me a while to track down which records had caused the error. If you can easily spot which entries in your load file have produced a suboptimal result, corrections become much quicker. LibraryThing has a Power Edit mode that enables you to quickly add, remove and replace tags, so it is easy to get rid of your dummy tags when you are done.

If you are tagging books, and working on a Mac (as I do), Typinator is a great help: you can establish abbreviations for frequently used tags, and save many keystrokes (if you are using Windows, this won't help you).

Of course, now that I am loading old books into the system, I'm finding a lot that need cleaning or repairing, so my dining room table is covered in piles of books that need help. And I'm finding so many forgotten gems on the shelves.

The CueCat scanner arrived on Friday, and it works as advertised. I've tried it on my MacBook Pro and my iMac, and it does what it says it does. You do get a dialog box as the Mac recognises what it thinks is a keyboard, but you can just close that and move on. Give the CueCat about 15 seconds to get its act together when you first plug it on. Actual scanning technique seems to be the only issue, and I imagine that that will improve as I scan more books. What is scanned is, as the instructions say, gibberish:


that's the scanned barcode for Martin Gardner's "Mathematics, magic and mystery". Feed it into LibraryThing, and it returns the book's details, right down to the correct cover art.

You should be able to see a link to my LibraryThing collection on the left hand side of this page: I've loaded 2,798 books so far, and I have a long way to go....


Bookmark and Share