Biblioracle: New Internet Fount of Book Tips

by Deirdre Foley-Mendelssohn

As a child—an easily bored, semi-feral child without a TV—I spent a lot of time in the local bookstore. The store had a large children’s section, with rows and rows of chapter books that led out to a small café, but by the time I was eight or nine, I would peruse the stacks and come away with the distinct impression that I had read everything there. The only thing—or, rather, person—that stopped me from giving up and turning to some other sort of entertainment was the children’s bookseller, a short black-haired woman who had read everything and could, if I told her some books I liked, recommend a new one to me—inevitably a more obscure but equally good one—with seemingly magical accuracy, the way that other adults enjoy pulling quarters out of kids ears. It was astonishing.

The recent appearance of the Biblioracle www.metafilter.com/92310/Biblioracle had a similar effect on me. Offering to suggest a book based on one's last five reads, he was impressive in his range, perfunctory in his responses, and quick on the draw. (Though I did wonder whether the public nature of the enterprise altered the responses. Were these really the last five books people had read, or only the ones they wanted to openly confess to? It’s like having your admired English teacher ask you what you have been reading: you aren’t going to confess it was "The Carrie Diaries"). And the oracle was prophetic in more ways than one: he didn’t just recommend books, he also showed that something is lacking in the way we buy books now, namely that black-haired woman. The oracle was quickly overwhelmed by demand, and queries began to go unanswered. One reader had a suggestion of her own for those stranded: “What to do if you don't get your Biblioracle recommendation request in during the designated window? Go talk to a librarian. They've been doing this sort of thing for decades.”

Of course, friends recommend books—but usually based on their tastes, not yours, and then reading can feel onerous, like a book report crossed with a blind date. There are also algorithms and Web sites intended to provide this sort of service, but I find them unreliable and broad. The suggestions on Amazon, for instance, are limited—they tend to pound me over the head with new releases, analogizing books based primarily on sales rankings—and often odd (this morning, in a rise-and-shinish sort of mood, my Amazon site recommended that I might like pancake mix). I’m sure someone could build a better program—a sort of Pandora for books, with user-feedback creating an increasingly subtle set of recommendations—but the rate at which we read books is simply slower than the rate at which we plow through songs, and thus it would take a lot longer to develop a truly sensitive system. It’s one thing lost most in the age of the Internet, and it makes the Web seem like a place very far from that cozy children’s bookstore, and much closer to some of the fearsome landscapes described within it—in the most literal sense, bewildering.

Note: Biblioracle is still working, but closes to traffic at 12:00 noon coast time.

Read more: http://www.newyorker.com/online/blogs/books/2010/06/the-trouble-with-recommending-books.html#ixzz0qTQxpmrp