Google Book Search -- Fair Use It Is Not
Google's "Print" (Book Search) was the cover story in the December 2 issue of the Chronicle of Higher Education's Review. The article, like many others, has somehow managed to avoid the flaw in Google's legal reasoning.
Google's scanning/indexing project creates a "derivative" work. To create the index Google must first create an electronic copy of the original. It is the process of creating this electronic copy (i.e. scanning and OCR'ing) which is forbidden without the publisher's permission. ONLY the copyright holder has the right to CREATE a derivative work.
Google has somehow asserted that transforming a work into another medium, NOT altering any of the text, preserving the transformed copy in a lasting manner, and doing so for a commercial purpose is fair use. But, what they have done is create a derivative work -- and there is NO fair use for the creation of a derivative work. Sure they can make a fair use argument about the index -- but the dispute is not the index: it is the process of creating the index.
Harper Collins has the right idea. They will control the process of creating the index and then will give the index to whomever wants to use it. The process of creating a derivative work stays with the copyright holder.
Of course, if Google were to play ball with the publishers then much of this would be unnecessary. Most books in the last ten years were prepared with digital files and it is much easier to prepare an index from the digital file. But that would mean Google would have to give up having its own derivative work.
Why does Google want a derivative work of its own? Because copyright law asserts that if a derivative work is created with permission or under license -- OWNERSHIP of the DERIVATIVE work is vested with its creator -- i.e. Google.
Anyone want to bet that once Google has ownership rights it will vigorously protect them?
Now if Google were to give up ALL commercial rights to the project, give it and the necessary funding to a consortia of university libraries (such as the ones they are "working with") and then find a way to restrict the use of the index to educational ones (users must log in through a library, text is displayed for only limited times and is deliberately fuzzy except say for a floating box which allows you to read paragraphs, and the ability to print and read multiple pages is severely restricted), well that might be fair use .... but that is not the Google project.
See: Andrew Raff's roundup for more