Saturday, October 25, 2003

Searchable Books

The Internet is buzzing with the news that Amazon is going to allow people to search the full text of books (or at least some books), with certain restrictions to keep people from circumventing the book-buying process altogether.

This excellent Wired article is emblematic of the news coverage -- full of praise for Amazon's initiative and full of eager anticipation over the wonderous aid to research that will result.

I do not disagree that much research would be easier if one could search the full text of books as easily as one searches the web. (Better yet, if one could do Boolean searches.)

Yet I'm worried about one small thing: The value of serendipity. I've found -- and I'm sure I'm not alone -- that some of the best and most useful books that I've ever found are ones that I wasn't looking for in particular, and that didn't turn up on a computer search. When I head into the library with a list of the books that turned up on a computer search, I often realize that the computer turned up books that were not what I was looking for. Instead, some of the most useful books end up being those on the surrounding shelves. And I never would have found them unless I took the trouble to stand in front of a physical shelf of books all afternoon and flip through one or two hundred books on the general topic.

There are several reasons why searching books electronically won't be as helpful as some may think. First, electronic searches suffer from two extremes: Make the search too broad, or use too-common words, and you end up with a glut of resources that are mostly irrelevant. Make the search too narrow, and you miss most of the relevant resources. Most people will err on the side of searches that are too narrow, thereby missing many useful books

Second, when you are starting to research an unfamiliar topic, you often won't even know -- yet -- the very terms that you should be searching for. Thus, you'll miss many books simply because you don't know exactly what terms to feed into the computer.

Third, the English language has many words, and writers have vastly differing styles. Even if you are familiar with the subject already, a computer search might miss many relevant books because the authors used slightly different words to describe the subject you're researching.

So I worry that the realm of books will become Googlized. If the text of many or all books becomes searchable, people will think that once they've done a few searches, they have exhausted the knowledge that the world of books contains. And they won't bother to visit the library just to browse the shelves and see what else is out there. Instead, they will continue to gaze blankly at their computer monitors, oblivious to all the knowledge that lies just beyond their grasp.


Post a Comment

Subscribe to Post Comments [Atom]

<< Home