Fair Use is a construct in US copyright law that consists of four principles that a court must use in determining whether an accused use of content is infringing or "fair."  It is not an a priori "list of uses that are fair"; therefore it is not possible to construct a DRM system that allows for fair uses and blocks others.  Fair use decisions rely not only on how the principles are interpreted but also on assessing the user's intent, which is always a dicey proposition. 

At the end of the day, DRM technology is (or at least ought to be) about narrowing what one might call the "benefit of doubt zone" among uses of content.  At one extreme (no DRM), users are physically able to do what they want with content, and the owner must rely entirely on the legal system to prosecute uses it believes are illegal.  At the other extreme (strong DRM), technology makes it difficult for users to use content in any ways other than those that the content owner is comfortable with.  To be palatable to consumers, a DRM system may allow some uses that can lead to infringement, such as the ability to burn music to CDs in MP3 format.  Content owners have to decide how widely to trust users in such situations when determining how to design or configure DRM systems.

A handful of researchers have advanced ideas about how DRM systems can be designed to narrow the benefit of doubt zone and do a better -- if not perfect -- job of supporting fair use (or its equivalents in some non-US copyright laws); these include Mark Stefik of PARC, John Erickson of NetRights (now of HP Labs), and most recently, contributors to the Digital Media Project.  Sun Labs' approach to this problem is innovative, though still not a perfect solution.

The DReaM approach to fair use is described in the white paper Support for Fair Use with Project DReaM.  It incorporates two key ideas.  The more interesting is the concept of an anonymizer, which is a piece of functionality that sits between the client and the server.  The anonymizer logs all uses of content and reports them to the server but without any identifying information about the user or device involved.  The audit trail that it maintains, which does include personal information, can be used to resolve disputes over fair use that come up if a content owner objects to something.  As the white paper says, the anonymizer functionality must be trusted, meaning (among other things) that it must be impervious to substitution hacks while developing a reputation for not revealing personal information except under egregious circumstances.

The second key idea is that of a process for a user to obtain "fair use rights" under DReaM: a user requests fair use rights, then up pops a window asking the user about her intent -- review, education, parody, or "other."   This functionality is supported by DReaM's MMI ("Mother May I") technology, which is in essence a protocol for negotiating content usage rights that requires users to reveal details about their intended usage in exchange for more rights.

In other words, the scheme described in the white paper requires a user to request fair use (rather than merely use the content without asking), provide some indication of the type of use intended, obtain the content for that use anonymously, and have her personal information revealed to the content owner only if the latter takes the trouble to request it because it suspects misuse.  This scheme does narrow the benefit-of-doubt zone but requires users to take extra steps and to have some understanding of fair use concepts, both of which are drawbacks. 

The other Sun Labs DReaM white paper, DReaM-MMI Profile for Creative Commons Licenses, is the result of an attempt to get DRM and Creative Commons content licensing to coexist peacefully.  The white paper mostly describes how this is not possible because of "the spirit of Creative Commons."  Technological support for Creative Commons in the DReaM-MMI model is reduced to notifying content owners of certain types of uses specified under Creative Commons licenses and does not include any support for protecting content so that it is restricted to Creative Commons-licensed uses.

We have repeatedly said that despite all the anti-DRM rhetoric from Creative Commons, its licensing schemes can coexist peacefully with DRM -- just not in the way contemplated by Sun Labs.  The idea of coexistence is not to get DRM technology to enforce the existing Creative Commons terms (which include things like the right to redistribute with proper attribution and the right to share content under the same terms as it is being shared with you); it is to use DRM technology for rights other than those specified in Creative Commons licenses.  Among the growing number of Creative Commons startups, at least one of them -- CloakX -- appears to understand this. 

Creative Commons is primarily intended for individual content creators and not large commercial content owners, but its applications in commercial media are both intriguing and barely explored.  Sun Labs' exploration of Creative Commons' coexistence with DRM is helpful mainly as a way of understanding the limitations of that coexistence.

Sun Labs is just about the only R&D entity left on Earth, apart from the Digital Media Project, to be seriously working on designs for DRM technologies that balance the interests of content owners and the public, in the way originally intended by Stefik and other DRM researchers.  DReaM arose out of a DRM interoperability project that Sun did for the EU-based EURESCOM initiative back in 2003 and has been under development for the past couple of years. 

It is far from clear how or if DReaM has a path to market; Sun Labs, after all, is an R&D lab, not a product development organization.  And given its orientation, the DReaM project is one from which Sun's media industry sales and marketing functions may want to distance themselves.  Nevertheless, the work being done on DReaM has intrinsic intellectual value that ought to be encouraged.