When Jon Zittrain and I began building the Berkman Center two years ago, we committed ourselves to action as a guiding principle, on the premise that the subject of our study didn't yet exist--the Net itself was still evolving. In constructing the Berkman Center in cyberspace, we would inevitably confront the very problems we wished to explore: problems of copyright, fair use, authorship, and ownership on the Net. The challenge was to build a structure by which our students could learn to relate to the Net; and by this same structure, to learn about the modes of behavior, motivations, codes, and customs that are coming to define it.

The question of copyright on the Net is more than a legal one. Let me tell you a story about the MIT Shakespeare scholar who taught me how copyright impedes scholarship. He's excited about the recent revolution that has been taking place in Shakespeare scholarship due to the creation at MIT of a database containing all of the texts which precede Shakespeare. Not some texts, but all texts, meaning that Shakespeare can now be studied not as a separate work of art--an oeuvre standing alone--but in the context of his and other cultures. The proliferation of texts in the present day makes the idea seem nearly inconceivable, but as you go back in history the volume of text becomes much more manageable, especially when you consider that what we have is only what has survived.

The momentum behind the project is so great that MIT is continuing to build this database from the past into the present, adding in all texts. But as the MIT scholars move forward towards the present, they run into copyright, and they begin to omit texts. Given copyright law as it stands, there is no feasible way to clear rights for a project this comprehensive. So the texture of the database begins to open up, like swiss cheese with holes in it. The holes become larger and larger until they eventually merge, and the content of the database effectively evaporates.

This means a Faulkner scholar can't do what a Shakespeare scholar can. The electronic revolution in scholarship brings unequal benefits. How do authors benefit if their work is omitted from an online database which may become the new virtual, global public library? Should copyright law that was designed to protect authors in traditional print media become the de facto arbiter of what constitutes the online canon?

These are the kinds of questions Jon and I grappled with when began construction of a virtual university: the Berkman Center project "H20--Harvard 2.0". This is an open university, offering educational content for free. There is no advertising in the traditional sense, beyond that which follows the public broadcasting model. Our first offerings were two cybercourses taught by Law School professors: Arthur Miller's "Privacy in Cyberspace" and Terry Fisher's "Intellectual Property in Cyberspace."

Arthur prepared "Privacy in Cyberspace" with his team of research assistants quite on his own; no copyright issues arose. But Terry wanted to post examples of copyrighted materials on the web, in the course of discussing issues of intellectual property. He came to me with the question: shall we clear rights?

We had three choices. Option 1: put the course materials together just the way Terry believed they should be. Clear the rights, no matter what it took in terms of time and money.

Option 2: put together the course materials together but omit the copyrighted parts. Replace omitted materials with what's already in the public domain.

Option 3: Put the course materials together the way Terry believed they should be and post them on the web with a notice, right up front: if any copyright holder objects to the inclusion of any of their copyrighted material, let us know. We will remove it. As an educational, nonprofit entity, we claim fair use, meaning we will use, but not abuse.

While Terry and I were still mulling over the question, John Perry Barlow, EFF Co-founder and Berkman Fellow (and then a Fellow at KSG), asked me to teach in one of his classes. We put my dilemma before the class.

No votes for Option 1. One or two votes for Option 2, the argument being that the public domain is plenty rich, and if we wanted to get class materials up and accessible, it'd be best to avoid hassle. But why should the avoidance of hassle dictate the kind of teaching we do? Terry and I are Harvard professors; Harvard's motto is "veritas." Shouldn't there be a good reason for us to distort the basis upon which we're drawing conclusions? What happens if we go for Option 3? What happens if we feel we haven't got a choice?

To choose Option 1 would mean denying the exigencies of the medium. Choosing Option 2 would mean compromising our ideals as teachers. My argument was for Option 3, by necessity. But what would happen to Terry and me if we chose Option 3?

Barlow's students had an answer. The first obstacle, of course, would be Harvard. Harvard's General Counsel would call us to ask whether or not the Berkman Center understands the kind of exposure to Harvard that publishing copyrighted material on the web would bring. The second might be a press; perhaps even a university press. The third might be a rights group that protects software. They wouldn't be able to stay out of it. It would be too important for them; we could be setting a precedent.

As class continued, we saw exposed the host of giants looking over the shoulder of the Net; perhaps with a university press as their pawn, perhaps litigating alone. Suppose one of these giants took offense to Terry's posted material, appended note and all?

We could take it out, or we could fight. We would have to make a judgement based on the merits of the case. We would have to assess whether or not we were abusing rights and if so, desist--or at least mitigate our decision.

Whatever Terry and I decided, we would be acting not only in our own and the Berkman Center's interests, but in the interests of veritas, of Harvard and the Law School in particular. Isn't the whole idea here to teach about the law of this new space, and how one creates such law? Terry's cybercourse copyright dilemma presented us precisely with the material of a clinical, legal education in copyright on the Net. And if we were to approach the issue in this way, what would be the downside? Our students could write the briefs; it was entirely possible that judges would listen.

Terry's cybercourse posed us with the question of copyright on the Net, but copyright wasn't the first issue we confronted. The first was a question of ownership.

Months before we developed the Berkman Center cybercourses, Jon and I decided to build a website for the Torts class we co-teach at the Law School. For this purpose we employed Eric Liftin, web architect; he produced an initial design for a virtual classroom, which we went ahead to prototype. We also contacted Lot 11, a company offering a product, Nucleo, which seemed to express the open architecture and object orientation we were looking for, and gave them Eric's specs. Our goal was to lay down a template in both substance and form, in both content and code, on which global education could be built. This was no secret to Eric.

Lot 11 was working under deadline to finish the site when Eric came forward with his concerns. Eric had given Jon and me his design, and now that Lot 11 was actually building it, he'd begun to think about the reality of the template succeeding. Eric argued that as an architect who had designed a plan, he should be entitled to recompense on each replication of the plan. Jon and I agreed.

We called Lot 11. We told them that our architect wanted a percentage of any copies of the template they sold.

At the moment we called, Matt and his team at Lot 11 had been going full steam ahead to meet the deadlines that we'd given them. Our concern was to get code up and running so we could road test it before Torts class began. What do you think happened then?

Did Matt say, Option 1, "Sure, no problem, just tell me what percent Eric wants." Or Option 2, "I'm going to have to think about that, but we are in the middle of building the site under deadline. Let's postpone this discussion." Or Option 3, "I'd better call my lawyer. What do I know about rights? Who can I ask? What kind of money is Lot 11 going to make with this? What kind of money am I going to make with this?"

Option 3 comes closest to the truth. All work on the site stopped, and everyone involved began to worry about their rights. The pie didn't yet exist, but we all started thinking in terms of what slice we were entitled to. If a dictate of cyberspace is that whomever is first to market wins the game, then it seemed to me that whomever builds out into the space, as we were, should not be required to negotiate rights forward. Rather, as in film-producing, recompense should be on the back end.

Jon and I talked to Eric and explained to him what had happened. We did the same with Matt. We gave assurances to both parties that should the template succeed and be replicated, Jon and I would make certain that they would do well by it. This turned out to be good enough for both of them. However, we had already lost tempo, and it hurt us.

The current dilemma Jon and I are confronted with involves the further development of "H2O," the virtual university, and operating systems.

We're at the stage where specs are rapidly taking shape. Starting from the HBS model, Tom Feegle of a company called Net Response produced these specs, in the form of a document which has both assets and limitations; the specs are clear on the transactional elements, but have a ways to go in terms of educational architectural structure.

Simson Garfinkel, a Berkman Fellow, is taking Tom's draft to the next level, raising issues of community, security, and the fundamental technical question: what operating system do we use? Unix or NT? Originally, I had assumed the Berkman Center would build using Unix, expressing our commitment--so well articulated by Susan Rogers at HBS--to open architecture, to modularity, and a nonproprietary system.

My doubts began when I learned that Rogers runs her platform on NT. Paul Martin confirmed these doubts. "What difference does it make?," he asked, "It's like the telephone wires. Content is what counts." Charles Wong advised me to go with a proprietary company to run the operating system, because we could rely on them for consistency and support.

So when I forwarded Tom's specs to Simson, I indicated that NT would be fine with me. Within a day and before I'd heard back from Simson, I ran into H.T. Kung. He and Paul had been spending some time at Microsoft; there were real questions about NT. A day later, at a meeting for our Second International Harvard Conference on Internet & Society, H.T. Kung and Scott Bradner focused with me on the Unix/NT question.

NT is closed, at various levels. First, the source code: is it accessible? Second, even if the source code is accessible, will it stay accessible? Third, even if it stays accessible, can one play with it? My questions are grounded in my concern for the function of a virtual university. Our students should have access to the operating system and be capable of being students of it, of creating with it; of making new things and seeing those things in use. This is not possible with NT. The observation is important for industry as well as academia, because it is out of academia that students emerge. If they've been unable to develop thought, skill and imagination around the complex problems of operating systems, technology will be constrained. It is out of creativity that technology grows.

Simson and Bradner agree. Bradner points out that Linux may be the way to go.

We are still learning, still deliberating. Will you join us? As Director of the Berkman Center and dean of the virtual university being built, I invite your input.