Google Book Search disappoints the scholarly searcher

University of California pro­fes­sor Geoffrey Nunberg pub­lished a schol­arly look at Google’s book scan­ning and search project in the Chronicle of Higher Education a few days ago.

In the arti­cle, Nunberg points out many of the mis­takes in Google’s library, such as:

• Listing 1899 as the pub­li­ca­tion date for books like Raymond Chandler’s Killer in the Rain, Robert Shelton’s biog­ra­phy of Bob Dylan and Stephen King’s Christine, to name a few.
• Searching for the word Internet in books pub­lished before 1950 pro­duced more than 500 results.
• Searching for Medicare in books pub­lished before 1950 got almost 1,600 hits.
• An entry for the book The Mosaic Navigator: The Essential Guide to the Internet Interface was appar­ently writ­ten in 1939 by Sigmund Freud and Katherine Jones.

Nunberg thinks that Google got into lit­er­ary water a lit­tle over its head when it took on this project:

I have the sense that a lot of the ini­tial prob­lems are due to Google’s slightly clue­less fum­bling as it tried mas­ter a domain that turned out to be a lot more com­plex than the com­pany first real­ized. It’s clear that Google designed the sys­tem with­out giv­ing much thought to the need for reli­able meta­data. In fact, Google’s great achieve­ment as a Web search engine was to demon­strate how easy it could be to locate use­ful infor­ma­tion with­out attend­ing to meta­data or resort­ing to Yahoo-like schemes of clas­si­fi­ca­tion. But books aren’t sim­ply vehi­cles for com­mu­ni­cat­ing infor­ma­tion, and man­ag­ing a vast library col­lec­tion requires dif­fer­ent skills, approaches, and data than those that enabled Google to dom­i­nate Web searching.

In other words, it’s just pos­si­ble that clas­si­fy­ing books to make them use­ful and search­able might be a lit­tle more com­pli­cated than run­ning an engine for com­mon searches, like the lyrics of a song or the dates of a war.

Have you used Google’s book search and found it lack­ing? Do you have hor­ror sto­ries about try­ing to find infor­ma­tion only to come up with man­gled results? Let me know your hor­ror sto­ries (or even your praise) in the comments.