Dan Cohen recently delivered a talk on what scholars want from a digital library at a Digital Public Library of America meeting. Reading his observations of scholars and their research habits forced me to reflect on my own research habits and my frequent reliance on serendipity in the stacks.
Cohen’s first point is that scholars want reliable metadata about books and other objects in the digital library. After all, we must assume that the information associated with a text is accurate before we can begin to trust the content of the text.
His second point is that most digital archives of scholarly material remove the element of serendipity–accidentally stumbling upon a source while searching for something else or discovering that a book had been mis-shelved but the text in its place fit in perfectly with your project–from much modern research.
Jonathan Rochkind’s first comment on Cohen’s article hints toward this, but I believe that most of my serendipitous finds in the library stacks have been because of metadata error. On several occasions, a mislabeled book sent me to shelves that I otherwise would never have visited only to lead me to a text that fit a need I didn’t yet know I had had. Too often, keywords were of no help in the initial searching but combing a general area once a good text was stumbled upon was usually fruitful.
Cohen and (I am sure) other participants in the meeting offered solutions to the problem of serendipity such as “more like this” links, “sample collections,” and even social connection spaces, and I believe all of those would help allow serendipity to run its course–especially if we allow some metadata to remain incorrect. Even if they’re digital, you never quite know what you’ll find in the stacks.