Google/Michigan/CHE

["Jim O'Donnell" <provost@georgetown.edu>]

See below on state of play with Michigan's Google-scanned books 
and Google's own service.  As the article points out, 
books.google.com now allows download of *some* 
long-in-public-domain books.  A couple of observations from 
practice there:

1.  The books now in Google from public domain scanning at the 
big libraries are very erratically present.  Individual volumes 
of multi-volume sets appear, but not the complete set.  Hence 
Thomas Bowdler's edition of Gibbon's Decline and Fall "for the 
use of families and young persons" is only partly there, likewise 
a still-used "History of Classical Scholarship" by J.E. Sandys -- 
only one of three volumes.

2.  It is possible that the complete sets are in fact there, but 
the underlying frustration in my first point is that the quality 
of the metadata is poor.  You can't in fact look for a set, or 
for a specific volume of a set.  The search and its results 
resemble those of an Amazon search engine, not a library 
catalogue.  If multiple editions of the same thing are there, you 
will have to click on each, look at pub. data and table of 
contents, and even then may have trouble being sure you see what 
you see.

3.  At this point it is wildly impossible to predict what will be 
there and what won't.  Not just a question of sets, but if you 
search by author, the impression given is that of a well-stocked 
antiquarian book store -- erratic, odd lots, famous work of X 
missing but minor works present, etc.

Summary impression (not surprising):  Google is not a library, 
and libraries still add substantial value.

Jim O'Donnell
Georgetown U.

______________

http://chronicle.com/cgi2-bin/printable.cgi?article=http://chronicle.com/free/2006/08/2006083101t.htm

Thursday, August 31, 2006

U. of Michigan Adds Books Digitized by Google to Online Catalog, 
but Limits Use of Some

By JEFFREY R. YOUNG

As it works with Google to scan nearly all the books on its 
shelves, the University of Michigan at Ann Arbor has decided not 
to make full-text versions of copyrighted books available online, 
even to on-campus users.

The university has upgraded its online card catalog to include 
full-text electronic copies of books that have been scanned as 
part of its controversial partnership with Google.

Users will be able to read the complete text of out-of-copyright 
works online. For those volumes, the university is making 
high-resolution images available for each page. That could be 
particularly valuable for researchers studying any material in 
the margins of the scanned books.

If a scanned book is still under copyright, though, users will 
not be able to read the digital copy. Instead, the card-catalog 
system will return a list of the pages that contain the search 
term and how many times the term appears on those pages. The 
reader will be directed to the library's stacks for the printed 
book.

Some observers had wondered whether the university might make 
full-text versions of copyrighted books available at on-campus 
computers, but Michigan officials ruled out that option early on. 
"We don't believe that fair use allows us to make that kind of 
access available to our user community," said John P. Wilkin, an 
associate university librarian.

The fair-use doctrine allows educators to reproduce a limited 
portion of copyrighted material for classroom use without seeking 
permission or paying royalties. But how far that principle 
applies to putting copyrighted materials online is an area of 
dispute.

Publishers have objected to Google's project, in which it is 
working with several major university libraries to scan books to 
add to its index. Although Google is making only short excerpts 
of copyrighted books available to users, publishers argue that no 
one can make digital copies of books -- even just for indexing -- 
without express permission from the copyright holders.

Michigan's contract with Google stipulates that a digital copy of 
each book scanned by the search-engine company will be given to 
the university for its own use. The university calls its 
digitized copies "MBooks."

Copy and Paste

A key difference between Michigan's digital books and Google's 
versions of the same texts is that Michigan makes it possible to 
copy and paste material from the pages. The pages offered on 
Google Book Search are image files that do not allow easy copying 
and pasting.

Last week Google added a new feature to its service, allowing 
users to download copies of some public-domain books to their 
personal computers. But those downloadable books, too, do not 
allow easy cutting and pasting.

The ability to easily lift blocks of texts for quotation would be 
a boon to scholars who are writing papers, but it could also make 
plagiarism easier. Mr. Wilkin said Michigan officials were not 
worried that their new service might somehow aggravate that 
problem. "Most of the resources we provide to students also make 
this sort of thing possible," he said.

The university's MBooks service was designed with academics in 
mind, he said. For one thing, the campus library is careful to 
make sure that all links to the digital texts remain constant, so 
that scholars can cite the digital books in their papers without 
worrying that the links will become obsolete.

Steven J. Bell, director of the library at Philadelphia 
University, said Michigan's new digital-book service could spur 
more scholars around the world to use interlibrary loans to 
request single pages or groups of pages from books held by 
Michigan. After all, if scholars can consult Michigan's online 
catalog to find out which pages contain the terms they are 
looking for, they might request just those pages rather than the 
entire book.

"It might really have some implications for the whole 
resource-sharing" aspect of library services, said Mr. Bell. "It 
really does simplify my ability to tap into that content from 
afar."

copyright 2006 Chronicle of Higher Education