Another Reason Why Google Isn't God

Google has, to the best of my knowledge, never claimed to be omniscient. While their oft repeated motto, "Don't be evil", might be seen as a claim of omnibenevolence, Google usually is straightforward about their interests: Reconciliation and redemption are outside of their purview, but data and dollars are right up their alley.

Thus it should be no surprise to the watching world that the Google Books project falls short of divine infallibility. (In spite of Google invoking the cadences of Genesis 1:1 in the project history.) This especially has been no surprise to the watching academic world which still has many mixed feelings about the benefits to be derived from the digitized corpus of the bibliosphere. Geoffrey Nunberg has written an excellent exposé of the shortcomings of Google Books when it comes to meeting the needs of researchers.

Of most interest to me as a librarian, though, are Mr. Nunberg's words of advice to librarians everywhere. He exhorts us to hope that Google "may be responsive to pressure from its university library partners—who weren't particularly attentive to questions of quality when they signed on with Google—particularly if they are urged (or if necessary, prodded) to make noise about shoddy metadata by the scholars whose interests they represent."

This blog post is only a leaf in the tossing forest of the Internet, but it is my effort to heed Geoffrey's advice and put in my $0.02. Our first response as librarians to digital books may not be embracing them, but, as librarians, we do have something to contribute when it comes to organizing them. Our catalogs (if I can persist in using such a quaint term) are not about the arranging of material volumes on physical shelves; they are about the organization of knowledge. Or so my Cataloging 101 course told me many years ago. Librarians are not divine, either, but neatness is next to godliness. Google Books is one horrific mess that needs a good cataloging. We already cataloged it once. Like housecleaning, though, it may need to be done again.

Bye-Bye Microsoft Book Search

The Chronicle recently posted an online article announcing the end of Microsoft's Live Book Search project.

Initial evidence indicates that Microsoft could not keep up with Google Book Search. I reported that Google had already topped the 2 million title mark in their book scanning project back in October; Microsoft is ending its project at 750,000 titles (plus 80 million journal articles). Given Google's commanding lead in the numbers race, it is difficult to imagine anyone else rivalling the Google Books collection for sheer quantity. The Open Content Alliance and others, though, may still be able to trump Google in terms of quality.

A Reply From Google

In response to my query about the availability of statistics for Google Book Search, Dan from the Google Book Search Team wrote:

Thank you for your email. As you may know, we've been able to include over 2 million books in Google Book Search so far. We hope to be able to release further information about the scale of Google Book Search in the future.

I hope that the Google Book Search Team can follow through on that promise. Having more data about the scope and content of the collection would be very useful.

More on Google & Microsoft Book Searches

I have been doing some more experimenting with both Google Book Search and Microsoft Live Book Search. I continue to be amazed at the speed with which these collections are growing, though neither Google nor Microsoft are publishing any statistics on the actual size of their databases, making evaluation difficult. (I have emailed Google about this lack of statistical data -- if they send me any information back, I'll let you know.)

I have, though, added both Google Book Search and Microsoft Live Book Search to our list of available library databases, and I am adding them to my information literacy class to give students some tips on how to make use of these databases. The extent of the available resources in these two collections makes them too valuable to ignore.

As an experiment, I searched 30 titles from our collection in both Google and Microsoft that were published before 1923 and therefore are in the public domain. From this highly unscientific sample, I found the following:

Available full text in Google: 10
Available full text in Microsoft: 4
Limited access in Google:10
Not available in Google: 10
Not available in Microsoft: 26

(Of the 10 that had limited access in Google, 7 were recent reprints--therefore under copyright--but 3 were the original, pre-1923 texts and should have been available full text.)

This then got me to wondering about more recent books, so I searched another 30 titles from off our new book list, picking titles that were published at least a year ago in case brand new items would not have had time to find their way into the databases:

None were available full text (not too surprising)
None were available in any form in Microsoft
Limited access in Google: 7
Listed in Google, but no access: 15
Not in Google: 8

"Listed in Google, but no access" is what Google calls "No preview available". You can find the title by searching title or author information, but the contents of the book are not available even for key word searching, much less previewing. I had not been aware that "No preview available" meant that the book contents are not searchable, but my own efforts to search key words and phrases from some of our "No preview available" titles confirmed this. In spite of their We-have-the-right-to-scan-any-book assertions, Google for whatever reason is not providing any access (beyond a title listing) to a certain percentage of their scanned database.

[Excursus: I find searching in Microsoft Live Book Search to be frustrating. There is no advanced search to allow searches by title or author; putting phrases from the title in quotes seems to be the most efficient way to search Microsoft. The response time is also much slower than Google, especially when scrolling through the results. I also found it odd that I kept encountering listings in Microsoft that said "Book Removed -- This book is no longer available" with no information to identify what the original book had been. If it's not available, why bother to provide a "Book Removed" listing?]

One practical question comes to my mind after this experiment: There have been rumblings that Northern Seminary may sell our current campus and relocate. This has necessitated a lot of work on my part to evaluate our collection and design possible scenarios for what to hypothetically do with the library collection should it theoretically be relocated to any number of putative sites. Most of these possible scenarios involve off-site storage for a certain percentage of the collection. Identifying titles available in Google Books appears to be one way to select volumes for storage. Dare one download and archive one's own copy of the PDF file and discard the book altogether? What would you do, hypothetically?

Google Books -- Pros, Cons, and ?s

I continue to be amazed at the speed of the progress of both the Google Book Search and the Microsoft Live Book Search projects. Three years ago, who could have predicted it? Or, more precisely, with decades of predictions that some day someone was going to digitize the libraries of the world, who really expected it to happen this fast?

While neither Microsoft nor Google are disclosing their costs or total progress for their scanning projects, a March 11, 2007 article in the New York Times estimated that Google had already scanned over 1 million books at a cost of over $5 million. (History, Digitized (and Abridged), by Katie Hafner.)

Of course, just scanning a book does not guarantee that anyone will ever be able to read it online -- witness the recent posting on the American Historical Association's blog, Google Books: What's Not to Like? There are many errors resulting from the high speed with which the Google Book Search project is being done, errors that render the scanned books unreadable or unretrievable -- the written word reduced to mere digital static.

Nor do we yet have a pleasurable way to read an online book. Online books are great for keyword searching or reading brief excerpts, but, as been noted in many places, you can't take them into the bathtub with you yet. Companies like E Ink may be close to breaking that barrier. Take note, E Ink: We want our electronic paper to be capable of color and moving pictures, and we want it to be tubworthy. It is also possible that online books will leapfrog text altogether and become part of the audio-drenched universe accessible through our iPod earbuds -- Project Gutenberg is quickly building a large collection of freely accessible audio books. Again, there are no statistics available, but a quick check of their list of new audio books posted or updated in the past 24 hours reveals 18 titles, not too shabby a rate, but still a slug's slow progress compared to Google's frenetic scanning.

A professor of mine from my library school days, Michael Koenig, has written about the three stage process of social transformation due to the introduction of a new technology. According to the theory, there is an initial stage of introduction of a new technology, a second stage of adoption where the technology begins to become broadly accessible, and then a final, third stage where the technology changes not only how we do things but transforms the very things we do. (See "Entering State III: The Convergence of the Stage Hypotheses," Journal of the American Society for Information Science, 43, no. 3, 1992.) For all the problems with the Google and Microsoft book scanning projects, my sense is that we are being chivvied into a state III transformation when it comes to online books. We're already cross-eyed from our efforts to track our online journals; what will happen to our library systems when our books are online, too? On the strength of JSTOR, many schools have already discarded back runs of their serials. On the strength of Google, will we discard our old books, too?