Skip to:

Amazon and the failure of metadata

Despite(or perhaps because of) being the world's biggest online bookstore, Amazon is a nightmare to use when searching for well-known books. The problem of finding a book on Amazon that exists in several versions has been very succinctly stated by Jim O'Donnell on a recent library listserv ("The Book-buying morass"), describes his problems trying to locate an adequate copy of Joyce's Ulysses for his library.

 

When you get a given title on Amazon, it generally lets you choose among Kindle, hardcover, paperback, and sometimes audiobook versions of the same book.  Time after time on the Ulysses pages, you will be given that choice, but the three or four versions whose tabs appear on the same screen turn out, when you click on a tab, to be *completely* different editions.  What looked like a possible contender for the paperback choice offered a "hardcover" tab that linked to an out-of-print edition by a completely different publisher.

 

A further problem is what he describes as "a vast mishmash of dumped-to-digital e-books of dubious provenance, dumped-to-POD print books equally dubious". 

 

The problem looks to becoming increasingly unmangeable: you can't any longer see the wood for the trees. You have to resort to clever tactics to find the edition you require of a well-known title - for example, I was searching for an edition of The Count of Monte Cristo with notes, say the World's Classics edition. On Amazon, that edition could be buried several pages down behind lots of scanned rubbish. If the books were only accepted by Amazon with a minimal specification of metadata, we the users wouldn't be spending many hours trying to locate the edition we want. Imagine a print bookshop where every well-known title existed in at least 20 different versions, 18 of which were in some way defective.