http://www.cdlib.org/inside/projects/massdig/faq.html#content

What rights to the digitized content does UC have in the projects; will access be limited in any way?

All contracts specify that UC digital images will be available to the UC Libraries to download and manage. The UC Libraries’ digital copy is subject to certain rights and restrictions regarding use and distribution. The University of California’s use or ability to display the downloaded copies of the full text of all books is subject to the restrictions of copyright law. Full-text searching will be possible for all of the digitized books, but some scanned books will not be completely viewable due to copyright restrictions. Specifics include:

Google

* UC Libraries have the right to use the UC Libraries digital copy at the University’s sole discretion, subject to copyright law, as part of the services offered to University Library patrons (including all individuals and organizations served from the UC Libraries websites).
* UC Libraries must implement technological measures to restrict automated access by crawlers, robots, spiders etc. to the UC Libraries digital copy.
* UC Libraries may not permit downloading for commercial purposes.
* UC Libraries may not knowingly permit the automated downloading and redistribution of the UC Library digital copy by third parties. UC Libraries must develop methods for ensuring that substantial portions of the UC Libraries digital copies are not downloaded from the UC Libraries website or otherwise disseminated in bulk.
* UC Libraries are permitted to distribute no more than 10% of the UC Libraries digital copy to other libraries and educational institutions for non-commercial, research, scholarly, or academic purposes (but not any portion of image coordinates).
* UC Libraries are permitted to distribute all or any portion of public domain works contained in the UC Libraries digital copy (but not any portion of image coordinates) to other research libraries for use by those libraries’ authorized students, faculty, and staff for research, scholarly, or academic purposes.
* Image coordinates, which link words in the OCR’d full text to specific locations on the viewable page, may not be shared with any entity.

Microsoft

* UC Libraries may use, copy, transmit, distribute, perform, display and create derivative works of the UC Libraries digital copy of non-copyrighted works to enable users of the UC Libraries and the end users of the UC Libraries websites to access and use these digital copies for personal, research or educational purposes, including, but not limited to, searching, viewing, printing and downloading by the UC Libraries end users.
* UC Libraries must use reasonable efforts not to knowingly enable any third party to download or host the UC Libraries digital copy or access or copy the UC Libraries digital copy via automated crawlers, robots, spiders, or similar data mining or extraction methods.
* UC Libraries may not permit third-party use of UC Libraries digital copies for revenue generating purposes.
* UC Libraries may license for non-commercial purposes the UC Libraries digital copy of non-copyrighted works and/or derivative works to non-commercial institutions such as other universities, libraries and archives.

Open Content Alliance

* There are no restrictions on access or redistribution placed on the UC Libraries digital copy.

There seems no hope (the text isn't quite clear) that in near future links to digitized content would be available via MELVYL. I don't think that OCLC's local WorldCat pilot is available for non-UC-users.

Among the current Google partners, at present only the University of Michigan has exposed the mass digitized content through its OPAC, with significant effort (including development of an ancillary rights database).

The quoted text is from April 2008 but this is evidently wrong:

In April 2007 Harvard's Crimson announced that Harvard's HOLLIS has links to Google's digitized books:
http://archiv.twoday.net/stories/4088388/ (in German, July 2007)

Since September 10, 2007, NYPL is offering access via CATNYP to the digital copies of its collections scanned by Google.
http://archiv.twoday.net/stories/4455443/ (November 2007)

It's a shame that most Google library partners (including UC) don't catalogue the digitized books by Google (or MS/IA) in their OPAcs.

March 13, 2008 Google has released, quoting the Google Blog, "a new API that lets you link easily to any of our books. Web developers can use the Books Viewability API to quickly find out a book's viewability on Google Book Search and, in an automated fashion, embed a link to that book in Google Book Search on their own sites." There is a link to the OPAC of a small Public Library in Oregon. (Harvard's HOLLIS is also using the API.)

Google Book Search provides a standard linking format that allows developers to link to books using ISBNs, LCCNs, and OCLC numbers. And rich universities like UC aren't able to negotiate with Google that they could get a specific API with an additional parameter (library partner) that would allow to link only to the own digitized copies?