John Miedema

Is there One Big Library? Or are there many distinctive local libraries? I think both exist. I am in the early stages of designing a WordPress plugin that will allow people to easy insert book data (cover image, title, author, publisher, etc.) into their blog posts. For the first cut of the software, I will select one of the four big digital sources of book data: Amazon, LibraryThing, OpenLibrary, and WorldCat (which I will call libraries). In order to make the selection, I made the comparison illustrated in the table below. In the discussion following the table, I tell how I decided to go with OpenLibrary. No doubt some will disagree with my criteria and evaluations; I hope to hear from them.

¹ An unconfirmed statistic from Yahoo Answers, apparently from Internet Retailer Top 500 Report 2008, stated that Amazon has about 10 million SKUs on the Web, including books, DVDs, etc.
² This figure likely includes DVDs and other item types.

Discussion and Decision

I have marked in green the library that I judged to be strongest for each criterion. Based on a simple count, OpenLibrary and WorldCat win, but stand even(I have corrected an error in the table; see my response to Steven in the comments) WorldCat stands one up over OpenLibrary. However, in the end I still favoured OpenLibrary. Here is my thinking regarding all four libraries:

Many bloggers link to Amazon when talking about books. Amazon has many nifty features, but it ranked lowest by my overall criteria. I was surprised to learn how few items Amazon carries relative to the other sources (though the figure is not yet confirmed; see footnote 1).

LibraryThing would easily have been the winner if it was not a proprietary source. That mattered to me. LibraryThing is a fantastic service that I have used for two years; I am almost at the limit of books for my free membership, and I plan to finally buy my lifetime membership. Spalding is a great innovator and partner with libraries. But I have long felt that there has to be neutral source of book data on the web. It could be argued that user-generated book data is a neutral source. I thought about this at some length, but this and the other criteria ultimately favoured OpenLibrary and WorldCat.

Since the decision is part of a book blogger’s plug-in design, I used book blogging and technical considerations to settle things.

• WorldCat easily has more titles, but there are many titles Not in WorldCat; in OpenLibrary, I can add them; that’s important to a book blogger who often reviews indie titles. If I had picked WorldCat, I would need another source for titles not in libraries.
• WorldCat has professionally curated data, but the many eyes of the web promise to eventually correct user errors in OpenLibrary; for book blogging, a margin of error in the metadata is acceptable; it is supplementary information.
• The fact that OpenLibrary is open source counted strongly in its favour. Open source products tend to start slower but evolve stronger in the long-run, e.g., Linux. It also appeals to me as someone who links to tinker under the hood, if not rebuild the engine.
• OpenLibrary seems to have a few bugs yet. See comments in this post). Well, there are advantages to getting in at ground floor. (Update: Did I say gamma?)
• Update. Given the comment about social data, I should add that social data was not a priority in this decision. OpenLibrary is the only source that does not have it, but it is not immediately relevant to presenting book data in a blog post. It’s on my nice-to-have-but-not-necessary list.

I choose OpenLibrary for starters. No doubt I will include the others in later versions.

Index