MR. ONSRUD: My name is Harlan Onsrud, of the University of Maine, and I will serve as moderator of this session. Let me first introduce the panelists. Shelton Alexander, of the Pennsylvania State University, is going to be our rapporteur for this particular session. Prue Adler is assistant executive director for federal relations and information policy with the Association of Research Libraries. Eric Massant is director of government and industry affairs with Reed Elsevier, Inc., and we will rely on Prue and Eric as our primary policy experts for this panel.

We also have with us Kenneth Frazier, who is director of the University of Wisconsin Libraries. Next, we have Tim Foresman, director of the Spatial Analysis Laboratory at the University of Maryland; and Barry Glick, who is former president and chief executive officer of GeoSystems Global Corporation. Certainly we have people who are both creating and using spatial databases.

So, let's move right to the questions posed by the National Research Council (NRC) study committee to guide this breakout session (see Box 12.1). The process I would like to use here is to start by reading the question, just so we know what question we are on at the time. Then I will ask Prue Adler and Eric Massant for some brief statements on a question-by-question basis. Then we will go for responses from the rest of the panelists to their comments and then from anyone else in the audience as well.

Box 12.1

Issues for the Discussion Session on Promoting Access to and Use of Government Scientific and Technical Data—An Assessment of Legal and Policy Options. Assuming that government data will remain exempt from intellectual property rights (IPR) protection, (more...)

Question number 1: Assuming that government data will remain exempt from copyright or proposed database legislation protection, what should be the scope of that exemption for databases created outside of government, with government funding? I would like to make it clear that when the term “government” is used here, it refers to the federal government and federal agency policies. That is something that has not always been clear; we are not talking about state and local government, and perhaps we might return to what those particular policies might be under the database legislation a little bit later. Another point is that we will assume that government data will remain exempt from copyright or proposed database legislation. Is there anyone that disagrees with that assumption?

DR. ALEXANDER: One exception is the National Institute of Standards and Technology (NIST), which has a special legislated mandate through the Standard Reference Act.

MR. ONSRUD: Right, and we will get into some specific exemptions, such as the former Landsat Commercialization Act and things like that. As a general principle, I think that is where we are starting from here.

Let's use an illustrative example for this first question. The government grants a contract to a commercial firm, and that firm creates a database. That would be a contractual relationship, or a government agency might give a grant to a university. The university, in doing its research, creates a database in conjunction with that grant. So, what should be the scope of exemption?

MS. ADLER: Before I tackle that question, I just want to make one point. In the context of all these questions I want to show my bias, which is something along the lines that the noted intellectual baseball player, Yogi Berra, once said, which is that this is déjà vu all over again. When I look at some of these questions I think there are a lot of lessons that we have learned because we have lived through a lot of this already. A lot of these discussions have already occurred and a lot of policies are in place. For example, we have the Paperwork Reduction Act that gives us guidance, Office of Management and Budget (OMB) Circular A-130, as well as a number of other policies that we have to sort through many questions. I think particularly with some of the examples we heard yesterday about what is going on in the European Union, we should keep those in mind as we consider each of these questions.

In that vein, looking at the question vis-à-vis funding of government-procured data, say, in a contractual relationship or something along those lines, those data should still remain in the public domain because they were done with agency funding or on behalf of an agency mission. Therefore, that is something, as we have heard yesterday, that the taxpayers have supported and those data should be available to the different members of the public in the capacity that they need to use them, be it a public institution, a public library, or Reed Elsevier who wants to take the data and value-add to them as a product.

I believe part of what we want to keep in mind on all these questions, but on this issue in particular, is that you want to see the kind of robust information industry in the public and private sectors that we now enjoy, and you don't want to disrupt that.

MR. MASSANT: I agree with all of that. I do want to mention that I am standing in for Dan Duncan of the Software and Information Industry Association.

The information industry has been a long-standing supporter of open access to government information, which has been reflected in a lot of the papers and the testimony that has been given over the years. The exemption for government information and for information compiled under exclusive contract has been in all the United States database protection proposals from the start and has been a consistent difference between the European community and us.

One of the things I want to make clear is that what the proponents of database protection are trying to protect is value-added products in which the private sector takes government information and, on its own, adds value. Now, this value is very expensive. The investment put into creating these products is sometimes referred to as “sweat-of-the-brow investment” because it involves collecting the information, cleaning it up, standardizing it, making it easy to use, doing a lot of things that are not currently protected. This sweat-of-the-brow investment results in products that are not necessarily protected under copyright. So, that is an important distinction. We are very much for exemption for any government-funded information or information products or services created under exclusive government contracts being protected.

MR. ONSRUD: Comments from other panelists?

MR. FRAZIER: I would like to make a brief comment from the perspective of the University of Wisconsin. While there is a presumption here that government data will remain exempt, there is also a presumption implicit in all of the items under Question 1 that we will have database-protection legislation. In this, it resembles copyright extension, in that much of the debate was confined to how we can make copyright extension less bad, rather than whether we would have copyright extension at all.

My institution is of the view that we should not have federal legislation that would extend copyright-like protections to factual data. I only want at this point to address a few points. One thing frequently mentioned during this workshop is that we don't really know what a database is. As a matter of fact, we know very well what a database is or could be. Nearly all forms of information and knowledge can be contained in a database.

So, what is in a database? The answer is, what you have. A database will include music, all forms of sound, images—moving and otherwise—text of all kinds, numbers, symbols, software applications of all kinds. In short, we are going to have a new regime for all kinds of information if we create database protection.

The university is continually involved in creating derivative databases from factual information. We believe that there is a fundamental question as to whether or not it is appropriate at all to have copyright-like protections of facts. That is to say, and it hasn't been said at any point along the way, that there is a constitutional question as to whether or not the commerce clause can be used to trump the copyright clause in the U.S. Constitution. I hope fervently that at some time that issue will be raised.

We have also heard about unintended consequences of database protection. I would like to remark that we can see very clearly that there are intended consequences of database protection. Database protection legislation is intended to make users of information pay more, pay more often, and pay for a longer period of time for access to information.

It is the clear objective of the commercial database industry to control every use of information controlled in databases. This is explicitly reflected in the contracts that in many cases prohibit such simple traditions as interlibrary loan, limit the way that information could be used in distance learning, and even limit the extent to which authors can use their own work in teaching and research. Our view is that databases are already bristling with protections—technological protections, which will become more extensive in the future; licensing protections, which are aimed not at commercial misappropriation, but aimed at the use of information. We will see this legislation used in exactly the same way.

Finally, I want to observe how fundamentally reactionary this legislation is. We are now looking at legislation that essentially maintains control of information by those who already hold it, or who believe that they can continue to control information into the future. We are now looking at technological options, for example, for redistributing government information without using commercial and not-for-profit intermediaries. Now, this is clearly a bad thing if you are in the niche business of redistributing government information. The Government Printing Office is in the same situation. It is going to be possible to bypass certain forms of commercial packaging and redistribution. I would argue that that is a commercially constructive, creative, and useful part of the economic model for information distribution, and it should not be discouraged. In fact, if anything, the creation of valuable derivative databases should be encouraged.

MR. ONSRUD: Let's continue with the initial statements. Tim Foresman?

DR. FORESMAN: I don't have a lot to add to what Ken Frazier said. I would point out that in the definition of databases, I differ only slightly in that the life cycle of a database from a scientific perspective is one that goes through many phases.

One of the difficulties that the National Science Foundation (NSF) deals with, in promulgating some recent statements of getting data out, is whether the data have gone through proper quality-assurance procedures. When do they become a database versus the more nascent component of a database? Although this may seem trivial, it is what holds back many researchers from wanting to distribute the data.

It is the tug and pull involved with data managers and various scientists collecting the data, in that they don't want the data to go out because they're afraid often that the data won't be properly used or may be misused. You hear it also from state agencies that we don't want to distribute these scientific data, which are technical facts and collections of facts, because they might be misused.

So, the definition of data shouldn't be just passed by. Those data reports, services, those kinds of things need to be revisited, to help put context on the distribution policies. That is the main condition. One way, technically, that we can deal with that is certainly insisting on metadata (i.e., data about data). Metadata are a philosophy. Knowing that the form and format and standards for metadata as a technical reality will evolve, and just establishing, first, that metadata can account for the various life cycle components of a scientific database—if we get those two put together, then we are fine.

MR. GLICK: First, I feel compelled to respond to Ken Frazier's point. Suffice it to say, I totally disagree. I think the view of my institution is that some form of protection, reasonable protection, is explicitly needed for factual information.

I do agree with Ken that copyright protection should not be extended to factual information, although I also believe that some of the recent court cases defined that somewhat narrowly and made the bar too high for proving creativity in a work. I do believe some form of explicit protection for massive investments in databases is needed. I do believe that the legislation should prevent the kind of abuses that Ken was talking about and needs to be balanced.

To answer specifically the question about government data, I think I agree with the comments made earlier. Government data should be exempt from IPR protection. I think that is fairly clear. Someone brought up the question about commercial or not-for-profit organizations that build their databases around government data. How do you protect the public's rights to continue to get access to that public-domain portion of the data, while protecting the value-added information that Eric Massant referred to? I think that is an issue.

Other issues are subsidies, as opposed to cases where the government is actually paying someone to build the database. What about subsidies that enable companies to build databases and so forth, such as small business innovative research grants, cooperative research and development agreements (CRADA), and small business loans? There are these cases that get a little bit grayer and we may have to look at how to protect the public's rights and the fact that taxpayers have helped subsidize commercial or not-for-profit databases.

MR. ONSRUD: Anyone from the audience? I will open up the discussion on this first question to responses to these issues.

MR. MASSANT: I would just like to make a point in regard to one of Ken's comments. As publishers, we create our products essentially for the user, and we want the products to be used however the user needs to use them. The database protection is to prohibit unauthorized uses that are going to cause harm for the market of the database. That we are trying to affect or prohibit legitimate uses is not at all the case. I think it is an important point to make.

MR. DE GIUSTI: Peter DeGiusti, Information Industry Association. I take some exceptions to Professor Frazier's remarks. One of them is the definition of databases. It basically does have a very narrow definition, or a narrower definition than has been proposed for what a database is.

Second, he brought up the issue of copyright protection. That really isn't the case. Copyright establishes affirmative rights, which is really quite different than the regime in the bill. The regime in the bill is based on tort liability, the ability of somebody to bring a tort cause of action. So, an exclusive right basically gives somebody a right, whereas our bill actually prohibits conduct, which is quite different.

Third, you bring up the idea of intellectual property under the commerce clause. There is already a big part of intellectual property that is created under the commerce clause. Trademark is created under the commerce clause. So, any constitutional question would be analogous to what would be brought up under the commerce clause there.

Then you bring up the idea of different terms or conditions that providers request; for instance, prohibiting interlibrary loan or other types of things. Those are things that the market should be able to figure out. If a database provider is very restrictive on the terms and conditions in their agreements, another database compiler will probably come into the market to provide a product that has different terms and conditions. Libraries and universities will be then free to get their information from that provider, and the market would probably give pluses to that provider because obviously interlibrary loan is important not only to libraries but to universities and others. They would have a plus in the market. That is all I have to say.

MS. ADLER: In some ways I may look in part to Eric Massant because he sat through some of the Senate negotiations that you referred to on Title V.

I do take issue with some of your characterization of Title V. In part, I am reflecting statements of some of the attorneys from the other side, in terms of the definition trumping copyright to which you refer. In terms of the definition of database, as we worked through trying to understand how big was the universe of that definition, Jon Baumgarten and others used the example of articles in a law journal and believed that this clearly was a database. We used an example of an anthology that was clearly a database.

As we went through, with the exception of some software and interfaces, it was accepted around the table—which, needless to say gave us pause—that the definition included just about anything and everything, short of certain routing tables on the Internet. These were exempt, as were most aspects of software. Other than that, there was the general understanding around the table—which we didn't agree to, but there was the understanding—that indeed, this definition was not narrow as you characterized it.

Second, and I think very important, in terms of the notion of not trumping copyright, I would certainly disagree in a number of areas. First, because there are no exemptions that many believe are meaningful in the bill, particularly for research and education, there is nothing comparable to fair use. And second, when we say fair use, we are talking about the host of exemptions—functions or operations such as interlibrary loan, preservation, etc—that are in the Copyright Act. These are completely lacking in Title V.

In addition, as this came up in discussion yesterday, part of what was most disturbing about our Senate discussions in August, September, and October was the notion that perpetual protection would be afforded to producers of a database. That is very different than the Copyright Act. There is no perpetual protection for copyright. In the Senate's discussions the proponents agreed that mere maintenance of something on a server constituted investment; therefore, protection was continued beyond the 15-year term. That would trump copyright. It also trumps copyright on the notion of protection of facts, which is what Ken was describing. In copyright, as you know, you can protect expression, but not facts. In sum, I would disagree with your analysis of what Title V did vis-à-vis trumping copyright.

MR. DE GIUSTI: I agree. I think fair use and perpetual protection are two areas that even Congressman Coble has identified that need additional work, and I am sure that our association agrees and will be working with you and others on the opposite side to do that.

MR. ONSRUD: We are getting off the first topic, but let it run just a bit longer.

MR. MASSANT: Let me just comment on a couple of things. As to the protection of facts, the legislation that has been proposed in Title V protects the compilation of and the investment in that database. The facts themselves, anybody can go out and do exactly the same thing to. So, it was not an attempt to protect the facts themselves. Anyone can collect whatever information; someone else can go do exactly the same thing. I think that is a little bit different from what you were suggesting.

As far as the maintenance is concerned, the maintenance was not for limited activities. I sat in on many of the negotiations in the Senate, but the maintenance did not include just maintaining a database on the Web. Maintenance meant updating, revising the databases. I think this is something that was concluded near the end of negotiations. That was my understanding of it.

MS. ADLER: I would prefer that understanding. That is not what was discussed.

MR. MASSANT: As far as fair use, certainly I think we are open to having fair use analogous to copyright included in database protection. Suggesting there was no fair use at all in Title V is not entirely accurate. In fact, the whole idea behind database protection is to prevent harm, which is one of the four criteria in fair use, in fact, one that is used quite often by the courts to determine whether a use is fair or not.

There was also a substantiality element in database protection, which is analogous to fair use. Even though it didn't look exactly like fair use, there were fair-use type elements. Whether those elements in database protection look like fair use, I don't really think that the proponents of database protection have very much of a problem with that.

MS. ADLER: I used the word “meaningful,” and I think that is an important context. There is an exemption in Title V, but those of us in the academic research library community didn't believe that it was meaningful. That is not to say there isn't something in there, but the provision was circular and ineffective.

MR. MASSANT: That is usually interpreted by the courts.

MR. ONSRUD: I am tempted to jump into the fray here myself, but as moderator I am going to try to keep out of it. As a scientist, I know that there are many forms of data that I can't reconstruct. I can't go out and measure temperatures again. If I am challenging the work of some other scientist, I need to be able to get at that extensive data set. One should not be able to hide behind trade secret and other potential types of laws, whatever intellectual property law it is, to prevent this progression of science to continue.

I said that I wasn't going to jump into the discussion; I have many other examples, a whole bunch of others, but I will stop at that. I would like to get back to Question 1.

DR. ALEXANDER: Question 1b would be of some interest. When you have databases that are built partially by government funding, partially by other sources that are not government, and assuming that you have no protection for the part that is funded by the government, how do you decide what is available and what isn't? Would anyone like to comment on that? I leave it as an open question.

MS. ADLER: I believe that this is a very difficult question. It comes back to the fundamental issue that government funds in support of a government mission have been expended in support of creation of this database. For example, it could be some of the global change information and it is done partially with university funds, private-sector funds, and government funding. I believe that the human genome that we heard about yesterday is a perfect example of private and public and other data. Having access to that core data is critically important.

MR. ONSRUD: Generally, any data set you would have, if somebody has added some value to it, this modicum of expression of originality or creativity has been enough to protect en masse copying. Doing en masse copying essentially would be a violation of copyright law. However, one still could extract the government data set; one could pull out the government data. That is the existing law under the current situation.

MR. MASSANT: If I could comment on that, I think there is a major distinction between government information or a government product or database that is produced as a result of a government contract and a product or service done independently of a government contract. Our view would be that if there is even 50 percent government funding for that information the whole product should be exempt; it should be in the public domain. Where it is done independently and the private sector takes that information on their own and adds value to it, then that is a very different issue. There are examples of this.

If there is a product created under government contract in which there is government-funded information and private-sector information merged into an end product, if the government information is still available from a government source, then it should be obtained from that source. If it is no longer available from the government source, then the government information should be made available by the private sector, but not the private-sector information.

If the database is merged, which is often the case, in such a way that you can't distinguish what are government data and what are not, again, I maintain that if the database was created under government contract, it should be in the public domain. If it was not created under government contract, then it is a totally separate issue and the resulting product should get legal protection.

As Prue mentioned, the Paperwork Reduction Act has principles that would require open access to information that is funded by the government. The Paperwork Reduction Act hasn't always been fully enforced. I think the exemption for government information in database protection is consistent with the Paperwork Reduction Act.

I think some of these issues should be really dealt with in the context of some of the existing federal laws—the Paperwork Reduction Act, the Freedom of Information Act (FOIA), those types of things—to make sure that that information is in the public domain.

MR. FRAZIER: I want to return to the most fundamental question, and that is the ownership of facts, any substantial part, language of that type. I have been aching to say this all along. We have heard many references to Feist as an example of a problem. I invite everyone to look at Feist and the reasonableness of the finding in Feist because it absolutely has to do with the ownership of factual information conventionally arranged.

What enterprise could not claim that they have sweat on the brow maintaining any database of any kind? I can't imagine any enterprise that would be unable to clear that hurdle. If you look at Feist, you can see the pernicious possibilities of being able to restrict the creation of derivative databases, recalling Chris Overton's remarks that the creation of derivative databases, in many ways, equals progress in science today. That is what is going to be inhibited by this kind of legislation.

DR. RUMBLE: I don't understand what a fact is. Is a fact the fact that I made the measurement or that the temperature was 20 degrees? If I write a sentence in a paper saying that I used Methods 1, 2, 3, 4 to perform this experiment, is that a fact? Is the description of 1, 2, 3, 4 a fact? Consequently, what are the facts in any scientific paper. It is extremely difficult to make the judgement about what a fact is in those papers, unless they are protected.

MR. ONSRUD: Of course, the courts have been making that on a case-by-case decision, as to whether a particular extraction has involved extraction of facts and essentially taking the factual information and expressing it in a different form.

DR. RUMBLE: I am just reacting to what Ken said. I have read Feist very, very carefully and in it there are many concerns about what is a fact and things like that. I think there is a great deal of ambiguity of what scientific-facts are. To say that facts aren't copyrightable, but the expressions thereof are, I think there is a very subtle separation between the two and I think we are blowing a lot of smoke on the issue, quite frankly.

MR. DE GIUSTI: I think Professor Frazier raised one of the fundamental questions that should be addressed here—ownership and the question in Feist. If the defendant in Feist would have gone out and independently gathered those names, there would have been no cause of action and there wouldn't be under our bill either. It is not an ownership of the fact. It is basically a protection for the work that goes into collecting it in your database.

As was discussed in one of the talks yesterday, if an actor goes out and independently gathers that information, be it from a government source or by doing his or her own surveys or whatever, nobody has any cause with that person, even if I have a database that is identically the same in the information, not the selection or the arrangement.

The problem with Feist and the court cases that came after is that they continued to degrade the sweat of the brow, so there is no sweat-of-the-brow protection. There is kind of a thin protection on selection and arrangement of facts, but there is no real protection for all the investment and work that goes into that collection, arrangement, and selection.

MR. ONSRUD: We are really off the issue of government databases.

MS. ADLER: I just want to reinforce what Harlan said in response to what Mr. DeGiusti just mentioned. Oftentimes you can't go out and recreate the data; I can't go and count the number of cows in Vermont in 1976. I just can't do that. I can't economically go and compete with the New York Stock Exchange, because I cannot get those real-time quotes in the same amount of time that they have the ability to do.

What Harlan was saying, I believe, is particularly important. As we have all heard, standing on the shoulders of giants is what science does. The research community doesn't want to have to replicate research for cost and other reasons. They need, oftentimes, to build on what someone has done before.

I believe part of the concern is, that would not be permitted in the same way under H.R. 2652, the transformative use that we heard Chris Overton talk about yesterday. Actually, many of yesterday's panelists were very eloquent on the need to have this transformative use. It was a cost issue and, in some of the language, it is a question of whether that would in fact be permitted under Title V.

MR. ONSRUD: I think we have a clash of paradigms. We have a clash of a market paradigm, which is really based on, as we heard yesterday, trade secret and restricting access to data, versus the whole scientific paradigm, which has always been very much an open paradigm in building on the works of others and challenging the integrity and assumptions of each other. You can't prevent me from challenging your work, and you get it out there. The whole idea is to get the ideas out so that other people will read and build upon that over a period of time. I think these two paradigms are coming into a severe clash under this database legislation that is being proposed.

MS. CARROLL: Bonnie Carroll, Information International Associates. I think the question asked of this group is very difficult. There are so many specific instances, and it is hard to deal with that. A question that has come up in the past is whether the law of private commerce should apply. Congress legislated, like in the case of NIST, that they have copyrights on their data.

The question that I would like to hear a response from the panel about is that Congress, in its infinite wisdom, legislates in the public good. Are there cases where the public good is better served by giving the rights to data to a contractor? Is that a basis on which to start thinking about some of the decisions that we are making? It is a philosophical dilemma whether it is always the case that the public good will best be served by totally limiting independent access.

MS. ADLER: Do you have an example in mind?

MR. ONSRUD: I have an example if you don't. The intent with a CRADA is that an arrangement is made with a contractor and government, and some innovation is the result of research and development. Something innovative is really going to occur that otherwise wouldn't occur. The government wants this innovation to occur, so they will fund this group to do it. This group or company will do it because they will get the patent on the innovation and be able to capitalize on it. So, it is kind of a win-win situation.

MS. CARROLL: In fact, it happens quite frequently with copyrights in data.

MR. FRAZIER: There is no such thing as copyright in data.

MR. ONSRUD: The problem here is when CRADAs are used to give an exclusive arrangement not for research and development, but all of a sudden a company gets an exclusive arrangement in a data set. Then, in essence, they combine and rearrange government data and now provide that as a service out to the general public. Then there is this exclusive arrangement. Even when the CRADA releases, is there now a, what, 95-year copyright interest in that new product that has been created in that collaborative arrangement?

MS. ADLER: One example that is always used is Landsat, which was not a positive experience. Given the experience, Congress changed the law. A lot of people are debating whether it has changed enough. Landsat is an example of Congress legislating along the lines that we are discussing. There was a clear sense of harm to science, to the community, and Congress subsequently reconsidered and came up with a different policy, a more open policy.

The New York Stock Exchange is an example of a “collection of information” that is regulated. These are a limited number of databases where there are regulations that are superimposed on the use of whatever those databases or data happen to be. In this case, the Securities and Exchange Commission has specific regulations and public hearings, and the like.

I believe the guidance that we have today is the Paperwork Reduction Act, which says no exclusive licensing for privatization; correct?

MR. GLICK: Governments have limited resources and they have to prioritize where those resources are placed. If the creation of a database, for example, is not such a high priority that it passes that bar, but it is something that does serve a public good—and Landsat, I think, is an example—the question is, should the government totally fund that database itself. If the answer is no, but the government would like to see that get done for either public-policy reasons or internal agency use or whatever, then creating an incentive for the private sector, I think, makes sense as an option that the government could use, to get things done that it wants to have but doesn't want the burden to fall totally on taxpayers.

MR. MASSANT: I would like to comment on that. I think I do agree with that. I think agencies have missions, and if part of their mission is to disseminate information, then that should be funded by the agency. If it is not part of their mission, they shouldn't do it. I think it is important that they not give away or sell access rights to cover the costs of their mission.

I think as far as getting the private sector involved, they should create an environment, if they can't afford to do something that they think is important, that provides an incentive to the private sector to take those data, but hands off, as opposed to being involved in a contract or selling the data or an exclusive arrangement.

MR. ONSRUD: This is Question 1c, which asks, Under what terms and conditions should the databases created outside government with government funding be made available? What are the salient factors that need to be considered? Shelton, do you want to make a comment?

DR. ALEXANDER: With respect to the responsibility of transferring a data set to the private sector, there is perhaps one unintended consequence. Unless there is some requirement imposed that the data be kept more or less in perpetuity, then those data could just disappear when that commercial interest no longer finds it to their benefit to mess with them. Data collected by the government can't just be pitched out. They have to be preserved in the National Archives or some equivalent thing.

MR. MASSANT: Just a quick comment on that. The government appropriates and then stops appropriations and then the government funding for information programs is withdrawn; the private sector is just as likely, if there is an incentive, to continue to provide access to data. I think it cuts both ways.

MR. ONSRUD: As far as an underlying data set, I would say if it is in the public domain, it is far more likely to end up in a library and be retained there. If it is a database produced by a commercial entity and it is of commercial value, then that database would be far less likely to arrive at a library. Is that not true?

MR. MASSANT: That is true, but I am talking about support for an ongoing database program, which is just as likely to be defunded from one year to the next.

MS. ADLER: Part of the issue here is an important one that everyone is touching upon. When an economic incentive for a commercial partner stops, what new mechanism needs to be put into place?

You are saying that traditionally libraries have been the depository libraries of government information. Given the nature of dynamic databases and the rapidly changing industry, what kind of a system do we need to build into this new environment so that, when a Reed Elsevier, for example, determines that it is not in their economic interest to continue working on a database, there is some public-sector access over the long term. Currently there is no infrastructure that is willing to pick it up and provide access to that resource.

In part that is the role in life of libraries. I think part of the worry that we do have is that there is not a mechanism currently in place under discussion that will deal with some of these databases in this new environment.

MR. MASSANT: Certainly, we have always been very supportive of the depository library program. In fact, last year there was an effort to revise Title 44 of the U.S. Code, which included some conditions for the Superintendent of Documents to look into creating a system for preserving and providing long-term access to government information. If it is information that the private sector put together, obviously that doesn't fall under the depository library program. There was a limit of 15 years in the legislation that was proposed last year. It really isn't in perpetuity. So, again, in 15 years, I think it would go into the public domain.

MR. ONSRUD: We've just discussed Question 1c. The next question, 1d, just so we cover all the issues, is, Should there be any restrictions on who can access the government-funded databases and should discriminatory pricing be allowed? I would like to gather some comments on those two items. Are there general rules or principles that could be stated here?

MR. MASSANT: Just to start off, I would say that there shouldn't be any terms or conditions under which the access to government information should be restricted, except in those cases of national security or privacy. There was an interesting point made with regard to the U.S. Geological Survey information, in terms of endangered species, and I guess those are difficult calls. I don't think there should be any restrictions.

MR. ONSRUD: Except for the standard FOIA-type of exemptions, there shouldn't be any restrictions?

MR. MASSANT: Right.

MR. ONSRUD: Anybody else want to comment on that, expand on that? Are we in agreement?

MR. MASSANT: I would add one thing in terms of price discrimination, which is a little bit different. I think it is important to encourage the private sector to invest by not discriminating against the private sector in terms of providing access to the data. I don't think pricing should be inconsistent with or actually violate the Paperwork Reduction Act, and this is really a fundamental principle. I think government information should also be provided to the private sector according to what the Paperwork Reduction Act suggests—for the marginal cost, or the cost of dissemination.

MR. ONSRUD: Is there anyone here who believes that the commercial companies should be charged at a different rate for access to government information than the general public or not-for-profit sector?

MS. SCHIRO-ZAVELA: Right now we have laws written about this. OMB Circular A-130 calls for agencies to distribute data at marginal cost to fulfill needs or requests. There is other legislation that allows data centers to recoup some of their expenses. There is a lot of legislation and policies on the books that, while it may not require discriminatory pricing, does encourage it. They say, we are cutting your funding; you can recoup it in the sale of the database.

MS. CARROLL: Information today is almost a necessity of life, like food and other things. For such necessities, the government provides some kind of assistance like food stamps. There should also be some kind of safety net for people who can't afford information.

I think in some ways, as information becomes more and more critical, if it is not differential pricing, then there needs to be a way that people can have access. FOIA is one way to give people access, although FOIA assumes that the information is available. We have structures like the depository library system, which provide the access directly, and you always have to have that kind of access. There are needs of society that have to be served. So, if you can modify the safety net or modify the pricing in some other way, I think that would be acceptable.

MS. ADLER: I agree.

MR. ONSRUD: I believe that this panel generally agrees, as far as this first question overall, that we have had a government policy of open access to government information, which is a good policy. Part of the Copyright Act says government shall not impose copyright against U.S. citizens. If new database legislation were to be passed—and again, there is a challenge as to whether new database legislation is a good policy or not—there seems to be agreement that the same policies that were in effect relative to the government's relation to copyright should be transferred to government's relationship in this new database legislation environment.

Let's move on to Question 2. What should be the scope of the government exemption when government data and databases are incorporated into a nongovernment database and disseminated by nongovernment entities?

For example, we have a commercial company that takes government data and incorporates them into their database. The question is, Are those original data still considered to be in the public domain? Another scenario might be if a commercial company enters into a contract with the government to create and disseminate information for the government. The company is doing that on a contractual basis. One could also have a scenario where a university is incorporating government data in a database. They are now combined with other data, but they're still the original government data.

MS. ADLER: I was just saying to Eric that I think he, in fact, had already responded to that in the context that it can be a very complex issue if, indeed, the government data are merged with nongovernment data. Access to the underlying government data should be allowed—I know this is what we believe, and I think it is what Eric said as well.

MR. MASSANT: Absolutely, where not done independently of government funding.

MS. ADLER: One always should have access to that underlying government information. The difficulty comes in how to parse out the underlying government information, if the information is merged together.

Again, I would also raise the issue that came up as a side issue. Over the long haul, if there is no other place to have access to that government information—if, say, Reed Elsevier is the contractor that the government has asked to create that database, and it decides over the long haul that it is not something in their economic interest to maintain—what kind of an infrastructure can be put into place to make sure that there is long-term access to that government-funded information; not necessarily proprietary, but to that government information.

DR. FORESMAN: I would like to talk about the complexity of following through on that. On one level, it would be nice if, when mixed, the subset of the government data was placed in the archives and would be available over time. The reality is that, through university research, private interactions, and a variety of mechanisms, getting those data back to the government, more often than not, doesn't happen. The data are kept; they are sequestered in universities; they are in databases. You really have to identify that specific data set to obtain them.

There are numerous government agencies that have contracts where we know those data don't come back to the contracting officer or technical representatives. They might get a report, which says here are some data, or we have the data. So, when you start talking about safeguards, et cetera, it is just manifestly difficult.

MS. ADLER: I agree. I think that is exactly the problem. I am not even advocating that it has to go back to government. There could be other arrangements, not unlike some of the NASA-funded arrangements now with some of the satellite data, where other institutions will take responsibility for those data. There has to be some mechanism to identify where those data are housed and what kind of access one would have to them. One really doesn't want to have loss of that record, for a whole host of reasons.

MR. FRAZIER: My only comment is, this is illustrative of what a mess you create if you create a right in data. Right now, we have a very low threshold for originality and the arrangement of data, which makes a great deal more sense than to create a right in the ownership of the facts.

MR. MASSANT: I just would like to go back to the issue of existing laws requiring that government information be made available to the public. I think in terms of when these contracts are set up, really, the government contracting official should include in the contract some condition that requires that the underlying government information be made available. I think that is very important.

MR. ONSRUD: Shouldn't that be left up to each individual contractor, or do it across the board?

MR. MASSANT: No, I think it should be consistent with the Paperwork Reduction Act. I realize that one of the problems is there are mechanisms like the CRADA that allow the agency to give rights, from what I understand, to some of the data that result from research. That, I think, is a problem and I think everything should be consistent with the paperwork-reduction approach.

MS. ADLER: I would add that it is not only an issue of consistency with the Paperwork Reduction Act. It is giving the Office of Management and Budget the authority to really make it happen. We have good legislation and good guidance in OMB Circular A-130 but not necessarily the ability to follow through and make it happen. I believe that is another part of the issue there.

MR. GLICK: I favor something extremely simple. Number one, if a private enterprise acquires government data and then adds any kind of value to those data, there shouldn't be any restriction on what that private company can do with the data; they can own that database, not the facts, but that database. Number two, there shouldn't be any requirement on that company to make available the government portion of those data, or to kind of put those aside. Maybe there are some examples that I can't think of, but why would you put that burden on the private enterprise when those data came from the government and exist in the government and those data probably don't exist any more at that private company in that original form?

MR. MASSANT: I agree with that. I think the government has a responsibility and it shouldn't be encouraged to abdicate that responsibility by turning to the private sector as a way to preserve and provide long-term access to the data. I think the private-sector firm that has invested in preserving the data should not be penalized later on for having had the foresight to preserve those data because the government didn't meet its responsibilities.

DR. SCHOOLMAN: Hack Schoolman, National Library of Medicine. I think that our lives would be much simpler and we would have a sense of knowledge and security in what was going on if we simply acknowledged the principle that government-funded data, no matter where they are, should be treated uniformly. Whatever exemptions exist for government data produced by the government, by government employees, or by government contractors, should be uniform throughout. Then we don't have any doubts as to what we are dealing with. It is a simple principle that I think would serve us well.

DR. FORESMAN: I concur, with very few exceptions—national security and privacy of individual, much like the census doesn't allow block-level information about individuals. All of those are well understood, so there would be very few exceptions.

I further encourage that keeping the simplicity under some rubric or terminology such as data democratization—once they are paid for by the taxpayers, the taxpayers should have access to them—should also be extended beyond the federal government. An artificial line has been drawn between federal and state and local governments. If you follow a lot of the funds that are actually involved in establishing many of the institutions and structures, such as who paid for the Internet, if you took an accounting approach to the federal dollars, you will find that they are not that easily separated. There are federal dollars mixed and there is a leadership role. I would presume that we not draw that line at the federal government, but try to promote the idea that taxpayers, citizens of the United States, if they pay for data—be they county or state—that they also get that same access. That is where there is a big division going on in the country right now, and that is where most of our problems are in the state of Maryland. That is not being implemented and there is a great deal of disagreement between local and state agencies and federal agencies on the access to government data.

DR. ALEXANDER: There is one issue that I wanted to raise, particularly in the context of the government sponsored university research. Under the Freedom of Information Act or other laws requiring access to data, if university research were included, and the data are to be made available on demand, in most of those grants there is no money left to provide that service. So, there is a practical question of what can be done if that condition is placed on the grantee.

MS. ADLER: I think Shelton is referring to the data access provision, which was included in the 4,000-page omnibus appropriations bill that passed last October. There was a provision on access to data generated under federal grants or contracts. It was originally in the Treasury bill, and then it changed significantly when it became law. In this case it had to do with access to research funded by the Environmental Protection Agency having to do with air particulate matter. Industry requested access to the underlying data and a university refused to grant access. It would give the research findings, but not the data. What has ensued in the legislation is that all these federally funded data will have to be accessible via FOIA.

I believe that OMB Circular A-110 is out in the hall, and that it will be revised—a draft revision—within the next week or two by OMB to address this new requirement from Congress, namely, how to make federal research grantee data available through FOIA. It clearly has members of the research and science community concerned. For example, there is concern that data could be made publicly available through FOIA before they have been verified, and there are confidentiality issues. In some ways, confidentiality issues are less of a problem because FOIA deals with confidentiality. There are a number of concerns that the research community has in terms of when the data are released, who they are released to, whether they will be misused for political circumstances, and the like.

I believe that is some of what Shelton was referring to. Congressman Brown introduced a bill last week to overturn that law from last year, on behalf of the science and research community. You will be hearing a lot more about it over the months ahead, particularly in the next 2 weeks, when the draft regulations come out. The draft will be on OMB's home page. I am sure it will be on a lot of electronic mailing lists as well. There will be a 60-day comment period, and then you will be able to see a redraft, which could either be the final regulation from OMB or there will be another round commenting upon the comments. It is more of a stay-tuned situation, but obviously, it is integrally related to some of this discussion.

MR. ONSRUD: This is also a good lead-in to Question 2c, which is, What should be the terms and conditions under which government or government-funded data are published in scholarly publications and elsewhere?

One of the questions here relates to NSF's policy for funded research and publications. is that good policy and should that remain? The policy right now is that you get a grant from the government, you publish the journal article, and essentially, NSF claims no copyright in that database. It is essentially up to the author to impose a copyright on it.

This raises an interesting issue here because with many of the journals, as well as part of the electronic journal environment, one is then often able to link directly to the data that one used in the research. Of course, those data may have been funded by the federal government, which is one way to get access to them, in this particular instance. The commercial publishers might also have an interest in this, because they are providing the commercial electronic journals in these particular instances. Would that apply to the data set or just all intellectual works created under grants?

MS. ADLER: We heard yesterday it is the research data.

MR. ONSRUD: So, it is the data, then.

MS. ADLER: Agencies currently have two choices when they make a grant. They can accept the grant research results, or they can waive research results when the grant is completed. One discussion that has been under way with some of the congressional committees and elsewhere is that the ability to waive would change. So, agencies would have to maintain access—this was in an Office of Science and Technology Policy paper—to all the federally funded research. In order for someone to be able to make an FOIA request, they would need to be able to go to the NSF, make an FOIA request of the agency, which means that the repository would have to reside within each of the agencies. That would also work in the model that you are talking about, that another researcher should be able to go and get those raw data underlying that scholarly journal article.

DR. SCHOOLMAN: If I can make one point, the published data in the scenario you gave, under the circumstances of the electronic journal as opposed to the print journal, the electronic journal would, under the definitions that are given in the database protection activity, be a database. Therefore, the extraction of data from the electronic journal would be prohibited under various conditions as opposed to the extraction of the data from the print journal, which would be in no way restricted.

MR. MASSANT: I think one distinction is that the data set itself, if it has been collected as part of a grant, would be exempt, essentially, under database protection. The article is somewhat similar to a value-added product. If the government has paid for the article to be written, if that is part of the condition of the grant, then neither should it be protected nor should the data in it. As with Elsevier Science, if an author writes an article that is based on the research that resulted from the grant, then that article should be protected. I think in terms of accessing the data that resulted from that research, those would still be in the public domain. I think that is an important distinction.

MS. ADLER: Can I just add one other thing to what you were saying? Whether we agree with it or not, Title V is not exclusive to electronic data, it includes print and electronic.

DR. SCHOOLMAN: I understand.

MS. ADLER: You are concurring with me. So, indeed, the print journal would have the same restrictions as electronic.

DR. SCHOOLMAN: There would be much more difficulty with the print journal constituting a database than there would if it were in electronic form.

MS. ADLER: I would hope that would be the case. Based on some of the discussions, I am not sure that is the case.

MR. MASSANT: I just would like to clarify. As far as data in an article, that is really a different situation. The legislation attempts to prohibit harm to the publisher. Any scientific publisher selling articles wants the users to make use of those data, but not in a competitive way. I don't see how someone extracting a few pieces of data from the article is going to have any effect whatsoever on the market for that article.

There is no issue here, I think, in terms of whether the data in an article are protected. I mean, from what I understand, a lot of these articles do not include data sets. There is a major distinction between the article, which maybe mentions a few elements from a data set, as opposed to including the whole data set.

MR. FRAZIER: I think it is very, very important to keep in mind that in the future, increasingly, data sets will be linked to articles. As this legislation is being proposed, as this concept of database protection is being proposed, there is no exemption built in. It will, indeed, be possible to claim the effects of government-sponsored research, because the data are going to be linked intrinsically due to electronic publication.

MS. ADLER: Let me just underscore that point in the context of some of the principles Justin Hughes talked about yesterday in the Department of Commerce letter vis-à-vis H.R. 2652. (See August 4, 1998, letter from Andrew J. Pincus, general counsel, U.S. Department of Commerce, to Senator Orrin G. Hatch, chair of the Senate Committee on the Judiciary.) One of the primary concerns was that the way that government information was handled was inadequate. The Department of Justice memorandum was also concerned along those lines. (See July 28, 1998, memo from William Michael Treanor, deputy assistant attorney, Office of Legal Counsel, Department of Justice, to William P. Marshall, associate White House council, regarding the constitutional concerns raised by the Collections of Information Antipiracy Act, H.R. 2652.) It also was concerned not only about constitutionality issues, but also about the access to government information. That remains an issue along the lines that Ken is talking about.

MR. ONSRUD: One thing that could be part of this legislation could be the safety net, which we referred to before, for the scientific community, as far as the right to read. All of these are intellectual works regardless of whether they are a data set, and how you read the data may be done by various methods, whether you create software or perform some other manipulation. If the library subscribed to a particular electronic journal, who in the scientific community should actually be able to read it and generate new products without having to pay royalties to use this data set? Who can construct new data sets, reconstruct, and combine with others without having to pay royalties back? I don't think that is clear to me. I think it is clear the other way. My assumption is that there would be royalties set up in that situation under the proposed database legislation.

MR. MASSANT: I am not as familiar with grants. Again, if data sets are generated as a result of government funding, they would not be protected under database protection. I mean, they clearly wouldn't be. That means that every scientist could take that data set and do whatever they want.

MR. ONSRUD: Why wouldn't it be?

MR. MASSANT: Our view would be that if there is any government funding for the creation of that data set, then it shouldn't be protected under the database protection—state, local, international, whatever it may be. I realize there are other laws that affect this and that some of those laws allow that kind of ownership. That we view as a problem.

MS. ADLER: I think Justin Hughes laid out the position yesterday very clearly, where he was describing it in terms of paying for it once. Then that information should become available. It is not unlike what you were saying in terms of a very simple principle.

MR. ONSRUD: That principle is close to the status quo in the federal sector, but it is not the status quo necessarily at the state level, for certain scientific data. It does vary across the disciplines as well. Any other comments on this?

I guess this relates to Question 1c that I asked the question about. NSF has a policy of leaving copyright ownership in the hands of the grantee. As such, Is there concurrence that we should maintain the status quo and perhaps have the same position in every government agency?

DR. SCHOOLMAN: May I point out that I think most government agencies have that policy, but they also have a regulation that gives the government the right to use this material, free of royalty, in whatever way it chooses. It is a limited transfer of copyright being made, even though publishers don't acknowledge it.

MR. DE GIUSTI: I don't believe it does include redistribution outside government purposes.

MS. ADLER: Depending on how this FOIA OMB Circular A-110 discussion plays out, the situation could change. I think that is the important dynamic right there. If, in fact, it would shift to the government agencies having to be responsible for housing data generated under these grants, which they don't have to do now, and an FOIA request came in, that would be outside the government using it for its own purposes at that point. Again, that is really up in the air.

MR. ONSRUD: Normally, if the agency doesn't have the record, they don't have to expose it.

MS. ESANU: Isn't one of the issues also that the principal investigators would have to provide the data, but they would not receive additional funding to provide the data?

MS. ADLER: Yes, absolutely. It is completely up in the air how this would work and whether the principal investigator would have to maintain the data over the long haul, as you are mentioning, or whether an agency would have to. The agencies are certainly not set up today, and would require additional funding as well to maintain and house all these data for however long. So, it raises a whole slew of different issues.

MS. ESANU: Of course, it siphons funds away from data creation.

DR. FORESMAN: If the data are without the attendant metadata, the data may be worthless. If I don't put in the references, you can't use the data. You have got bits and bytes. There are more nuances in there. In a grant, if you are not given the extra money to develop adequate documentation for your data set or the extra money to provide for distribution, then you are stuck in a situation of assuming responsibility because it wasn't paid for.

MR. ONSRUD: Okay, let's go on to the next question, which asks, Can the federal government use and archive data that are created by both not-for-profit or commercial entities? So, what should be the government's policy in obtaining these data from the commercial sector?

MS. ADLER: Some of what we heard yesterday from the genomics data panel was illustrative of that. You have public and proprietary data in the human genome database, and you have public and private people dipping in and out on a regular basis, which makes the database as robust and useful and effective as it is. Is that the kind of example that you are thinking of?

MR. ONSRUD: Yes, I think that is what was intended. I didn't write these questions, but I think that is the general feeling here.

MR. GLICK: I think the grassroots feeling in industry is that, once data are delivered to the federal government, they can't be protected adequately. With some minor exceptions, for example, in contracts where there are trade secrets and confidential information in a legal sense, I think there is some confidence that those get protected. In general, many companies either refuse to license their databases to the federal government or try to put in restrictive clauses. They are concerned with the possibility, either through FOIA or some other way, that their proprietary information will leak out and they will lose economic value potentially because of that.

MR. ONSRUD: One of the exceptions to FOIA is proprietary information. If the agency feels that they should not provide it, by law they are allowed to protect it. How well they protect it is not clear.

MR. MASSANT: We support strongly that any proprietary data provided to the government should not then be redisseminated. They should be protected. That is a very important condition. Otherwise, the private sector wouldn't sell data to the government, if they were at risk of that information being released and put into the public domain.

MS. ADLER: There are always yes, buts. Two different databases come to mind where that doesn't necessarily fit, given the nature of how this is framed. The Securities and Exchange Commission data are all generated by the private sector, but they are generated to meet a government mandate, which is then made publicly accessible. They technically could be considered proprietary data of each of those individual corporations.

Another example is the toxic-release inventory. Many, companies did not want that information publicly available. Again, those are technically proprietary data that have to be made publicly available.

MR. GLICK: I think those are special conditions as opposed to data that have been published to sell in the market. They are a different type of information compared to the SEC or the toxic requirements.

There is an interesting case at the Census Bureau where, in order to prepare for the census, the Bureau actually licenses map data from publishers and then incorporates them into their own public-domain database, which is provided as any public-domain database; essentially, free of charge or for reproduction costs. The requirement a lot of people ignore is that the information still carries the copyright of the publisher, even though there is no payment to those publishers.

As far as I know, it seems to work to the satisfaction of the publishers, but there is embedded private-sector data in a public domain database with copyright still protected. It really is supposed to prevent a third party, then, from pulling out the copyright portions of the data and using them to their own benefit, even though the entire database is public domain.

MR. ONSRUD: How does one identify the creative elements of that copyrighted work in that instance?

MR. GLICK: The creative elements are almost impossible to find. If, for example, you draw a map of Birmingham, Alabama, and that was one location where a private-sector database was used, you will see a copyright notice appear, either on print maps or, if it is done electronically, if it is done properly, you will see a copyright notice that there is no way to tell what elements of that database came from a publisher. My understanding is that.

DR. SCHOOLMAN: Isn't there a simple answer, that basically says that except for the restrictions imposed in the agreement by which the material was acquired, the government can do all those things?

MR. ONSRUD: I think it is already there. I think there is a status quo relationship in the government policy even before we had these database legislation discussions.

MR. MASSANT: As far as the elements that are laid out here, as I mentioned, I think fair use, or terms analogous to fair use in database protection, would be generally acceptable. The suggestions here go far beyond fair use. For instance, one of the things that is mentioned is being able to integrate data into multiple-source data products or databases. That would be acceptable if the resulting data product or database was for purposes of the user only. Of course, if we are selling any type of database, we would expect the user to be able to do things like that with it. What would be very, very problematic would be if that database in turn was then disseminated in such a way that there was no need for anybody in the market to purchase the original database.

I think the focus here is on the harm issue. If some of these things would cause harm to a database, that would clearly be a problem. Of course, ultimately nobody would invest in a database where they would sustain that kind of harm, and ultimately the user would suffer if those databases weren't produced.

MR. ONSRUD: You are saying that if several commercial entities delivered data under some arrangement, and a government agency took and combined these data sets or did some other compilation, and then made that available, even though in a copyright environment that would be a new creative work, the argument could be made that that would harm the market for a commercial product by the government making this new arrangement available? You see that as an inappropriate action?

MR. MASSANT: Absolutely, if it caused harm to the original database, then that would be a problem.

MS. ADLER: You mean, the market for the database.

MR. MASSANT: The market for the database.

MR. GLICK: I don't see why the government would be considered any different than any other licenser of the database, and license conditions would deal with all those issues. Except for FOIA exposure potentially, I don't see the government obtaining any automatic rights when they execute an agreement to acquire a database that is any different from any private-sector or not-for-profit customer.

DR. SCHOOLMAN: I would point out that you do precisely what you just indicated. In the Unified Medical Language System (UMLS), the National Library of Medicine has a number of proprietary databases. We have an agreement and we have a license, not only with the proprietary owner, but with the users of the UMLS. The UMLS may be used for research purposes as it stands, using the proprietary database. Any commercial product that is developed by any of our licensees of the UMLS cannot include in any of the proprietary databases without negotiations with the proprietary owner. It is a perfectly straightforward arrangement and the National Library of Medicine has no great difficulty in implementing it.

MR. ONSRUD: I was thinking of the scenario where such an agreement wasn't in place. The government just essentially did it on its own. It had access to this database, considered it to be raw data, created another product, and made it available. So, it is a little different scenario, I think.

MR. GLICK: How can the government take raw data that are a commercial product? You mean pirating them?

MR. ONSRUD: No. For example, you deliver a book to the library. A scientist will read 10 books, take the data out of them, create a new product, and make that available on the Web. The scientific community can do that on a regular basis. So, the question is, if we are going to continue this scientific paradigm of openness and ready use of each other's works, basically, you have a right to read and a right to access those materials and create new things and make them available. That new book may compete with other previous books, but we don't have to deal with licenses.

MS. ADLER: I wish Chris Overton from the University of Pennsylvania were sitting here. Part of what this gets to is what we heard from him yesterday. He takes chunks, as he used the term, of data from a whole slew of different sources, public and private. What he is increasingly experiencing is that this one database that is crucial to his work, which is an excellent database called SWISS-PROT, has put new restrictions on it. That includes downstream restrictions in terms of what can then happen, once his new database is created.

Thinking of it in those terms may be helpful in the sense that we have a specific example of what some of the European Union (E.U.)-type restrictions mean as you combine data from the different sources into creating, in his case, probably a government-funded database—not a government-generated, but a government-funded, database—that then has embedded in it license restrictions from one of the sources. Increasingly, what will that mean if 5 out of the 57 sources of data mining have different restrictions, each calling upon the user to do something else with these restrictions?

MR. MASSANT: In the case of SWISS-PROT, is it the Swiss government that is claiming rights on this database?

MS. ADLER: There was some question. It was a not-for-profit, but I think with the blessing of the government, was the description.

MS. ESANU: There are funding issues with that database as well. It wasn't necessarily the E.U. Database Directive that spearheaded these restrictions, but there were also some funding issues.

MR. MASSANT: I think this does raise an important issue in terms of the international issue, and Justin Hughes mentioned this yesterday. For instance, the United Kingdom claims a crown copyright and, of course, they would claim the same kind of right on databases. This may be an opportunity, as we go forward at some point, hopefully, in the World Intellectual Property Organization with discussions for concluding a database protection treaty, that the United States will be pushing for open access to government information and exemptions from protection for government information. That may be a way to get other countries to open up their government information. Certainly, as was mentioned yesterday, the countries that are not part of the European Union have not signed up or do not have to abide by the E.U. Directive. So, there is a possibility there of getting other countries to have a more open type of policy in terms of government information.

DR. SCHOOLMAN: With regard to SWISS-PROT, I don't think we should assume that the market will always pay. I think we should address another position, that the marketplace will work things out most of the time. In the instance of SWISS-PROT, I think that it will. The type of database that was alluded to by Chris is not competitive with the SWISS-PROT database. The type of restrictions that SWISS-PROT is proposing is costing them a customer eventually. Eventually, scientists will figure out a way to get what they need without anything from SWISS-PROT.

If we try to solve every one of these anticipated inconveniences in the marketplace, by some rule or regulation, I think we are doing a disservice. We are much better off letting the marketplace work these things out in all those instances, except where there is clear evidence of market failure.

MR. MASSANT: I think that is a real important point and I agree. Currently, SWISS-PROT provides an option. If the market feels that they are charging too much or that they are too restrictive, then the market can turn to an alternative. At least they have an option. If there wasn't an option, they would have to come up with another approach. At least in this case, there is an option.

I am not agreeing with the Swiss government or whatever institution that is claiming these rights, but I think you are right, that there is an option, and if the market doesn't like it, they can go elsewhere.

MR. ONSRUD: Let's move on to Question 4. If the government data are exempt from the scope of the proposed database legislation, should databases created in universities also be exempt?

MR. MASSANT: I really don't have an opinion on this, unless the database is funded by the government; consistent with what I am saying, it would seem that it should be exempted.

MR. FRAZIER: I think I can safely speak for a very large community of universities, and that is, we do not want rights in data. It will be a very big burden on universities. Much of what universities do today involves transformative use of data. The proposed legislation is very, very much unwelcome to the university community, as represented in the testimony during this conference.

MS. ADLER: When this issue came up, a very helpful House staffer helped me understand that many state laws demand that universities have open records, if you will, with the exception of obviously some privacy and confidentiality issues. I think Ohio has one of the more stringent laws in the country. Part of this discussion was trying to mirror some of the state laws that are already in place, according to some of the drafters.

DR. FORESMAN: I would like to offer an anecdote that tends to counter what Ken has said. I had to personally bring the federal regulations on the distribution of data to university lawyers to prevent the attempt to set up a profitability center on the data. I was relying on the standard contracts. So, I think that some universities are looking at economic enterprise centers as a way of the future. While I am not in disagreement with Professor Frazier, I am saying that I would not assume that in all universities there is a homogeneous environment.

MR. GLICK: If the professors at Harvard University can gain benefit from copyrighting scientific articles, results, whatever, and the University of Massachusetts professors can't do that, is that unfair, even if both were funded by the federal government? That is the situation that you were just saying you seem to think is okay.

MR. ONSRUD: Essentially this highlights a whole complexity of issues that will be raised if this database legislation, or something very similar, is passed. I would see that all the arguments, then, would be reiterated in every single state in the country. I can see the state legislatures saying, well, we should impose copyright now in our state universities, otherwise we will have a dichotomy in our state, between the private and state universities. What issues does that raise?

MS. ADLER: Not all the time. Think of some of the multiple digital library projects that span multiple universities and multiple principal investigators. Some will be in private institutions, some will be in public institutions, and some in the private sector. It is, more often than not, in programs like that, mandated that you have public and private participation. I think it gets murkier and murkier to have this kind of distinction.

MR. ONSRUD: In this proposed legislation, were state and local governments exempt from database protection?

MS. ADLER: It changed several times. I am trying to remember where it finally ended up.

MR. MASSANT: I think it was any government, no matter what kind it may be.

MS. ADLER: But there was a specific exemption for state universities.

MR. MASSANT: For educational institutions.

MS. ADLER: On the House side, it was a very strong exemption initially, courtesy of Barney Frank. Then Barney Frank again changed it so that it wasn't so explicit as an exemption for state universities for making information publicly available. So, it did change over time and I don't know how it read in Title V.

MR. MASSANT: I think that in Title V there was an exemption for state educational institutions, where they were exempted from the government information exemption.

MR. ONSRUD: So, the state office of geographic information systems (GIS), which is actually charging me now for certain services, their databases would not be protected, but the GIS operation at the university would be protected.

MR. MASSANT: Yes, I believe so.

MS. ADLER: Depending on your use. If you then turned around and used it for commercial purposes, absolutely not. I think the exemption was research and education purposes.

MR. MASSANT: It was both educational purposes and the educational status of the state institution as well. So, if it was a state government noneducational institution, it would not have protection, whatever the use was.

MR. ONSRUD: Comments from other members of the panel in response to this particular question? Again, this assumes that if you have database legislation, then what would be your response? Some people may resist or may challenge that assumption of legislation, but again, that is one of the assumptions that this set of questions started with.

I don't know that we have consensus here. Obviously, it was argued back and forth by those people debating this in both the House and the Senate. So, there are pretty strong arguments either way and major conflicts either way. There is a major conflict between the universities and a state office sharing information.

DR. FORESMAN: I do have a quick question. That is, any university whether state or private or public may get government grants, so that is a different issue. A state university is not necessarily 100 percent funded by the government. Couldn't there be some databases that come out of a state university but are not funded by government funding? I guess there was an assumption that all state university employees are essentially government employees. I think those are issues that may help determine what to do with this.

MR. ONSRUD: Most state university employees consider themselves to be employees of the state—at least I consider myself to be—since the budget of the university comes out the state budget. We have other sources of funding, of course, but a significant portion of that comes from the state. I guess the university may attempt to claim some ownership in your intellectual work at their option, whether it is a database or even copyright material.

Let's go to Question 5. What other important legal or policy issues should be considered in promoting access to and use of government scientific and technical data for the public interest?

DR. FORESMAN: One thing that hasn't been discussed, although it was brought up at various points, is, Should there be a policy that the government may not refuse access based on lack of funding? Often that is a reality at some levels. I don't recollect any policy, but I am not as familiar as some of the other panelists, where the lack of funding could be used as a barrier; where funding wasn't provided to the government when citizens tried to access the information. I mention this, because there are cases where people are generating new data sets that become of value to citizens that citizens weren't accessing before, but the funding to provide those data sets is not available. We are seeing this and the Web is promoting this. Should there be included in any language a point on the fact that funding cannot be used as an excuse to deny access? When the agencies are pursuing their missions and they have their budgets, could that be used as an excuse to delay or prohibit the access?

I just thought that should perhaps be looked at and it wasn't covered previously. We are acting like it doesn't cost anything to provide the data, yet we have talked about it many times.

MR. FRAZIER: A couple of other issues. One is that I think there is a very real possibility that ownership should be legally challenged, if it is brought into law. It is certainly contrary to several hundred years of American tradition.

Second, where is the compelling case for legal protection of data? I would argue on the basis of testimony that you have heard throughout this workshop that much of the academic community is strongly opposed to creating a right in data.

There is a causal issue, and that is, why would we want to impose such baggage, such impediments on the transforming uses of data, since we have many, many examples—some of them cited during this workshop—where data, factual information, have been transformed through creative new uses. There is an abundant history of such transformative use of data that takes place routinely within our research institutions. Why would we want to give up such a competitive advantage for American society in order to protect an extremely narrow base of private interest, which is really what is driving this legislation?

MS. ADLER: To follow on what Ken was saying, many of those points were brought out in the three memos that you heard Justin Hughes and others allude to yesterday, one from the Federal Trade Commission (FTC), one from the Department of Justice, and one from the Department of Commerce. (See the previously noted Department of Commerce and Department of Justice memoranda. See also September 28, 1998, letter from Robert Pitofsky, chairman of the Federal Trade Commission, to Congressman Tom Bliley, chairman of the House Committee on Commerce.) These memos teased out the legal issues and specifics to what Ken was just referring to.

One thing that came out in many of our discussions yesterday, which I found helpful and compelling, was so many of the industries that you heard from are innovating very quickly and very successfully, and are moving into being very robust industries. The FTC analysis of the legislation explicitly referred to it as being anti-innovative and anticompetitive.

Whatever path we go down, we should take to heart what Ken is saying, in the sense of ensuring that we don't so skew one community's—in this case, the academic community's—ability to operate in addition to other newer industry's ability to create new science and new industry. This is a challenge that we heard a lot about yesterday, that this legislation would disrupt.

In particular, what has been summarized a number of times is that in the government information arena, the current laws and policies stand very well and have served us very well. We should continue to let them do that, specific to government information.

MR. DE GIUSTI: I would just like to reiterate that I don't believe this legislation—at least that my association is doing—is to create a right in data. It is just to give somebody an ability to have some type of legal protection so that they do have incentives to collect data. It doesn't create a right in those data.

MS. ADLER: The legislation did speak to controlling all downstream uses of data. So, whether you call it a right in data or the ability to control any transformative use, the effect is the same. I believe we are talking semantics here, but important semantics.

MR. EISGRAU: Just a question to follow up on Paul DeGiusti's comment. In the legislation negotiations last year, there were all types of potential standards on which to base the protection. I may have heard something different than I heard last year. Did you mean to imply that if we were going to have protection, the trigger for that protection ought to be an action on somebody's part that eliminates the incentive for a database producer to produce or enhance a database?

MR. DE GIUSTI: No.

MR. MASSANT: Paul's point is an important one. The downstream uses that we are still talking about is where somebody has invested in creating a database, and downstream uses that will cause market harm to that database should be prohibited. That is not to say it would affect anyone from going back and collecting that same information. Those data themselves would not be locked up, which is a big difference.

As far as Ken's comment that there has been a long-time prohibition on the protection of facts, first of all, it doesn't protect facts. Second of all, as far as sweat-of-the-brow is concerned, up until 1991 there was protection under copyright for investments, and that had been going on for a long time.

Regarding the issue of whether there is a need for protection, there have been a number of lower court cases. If you look, for instance, at the Warren case [ Warren Publishing, Inc. v. Microdos Data Corp. , 52 F.3d 950 (11th Cir. 1995)], the plaintiffs clearly were harmed by someone taking their data and they had no way to protect the data. So, there clearly is evidence of a need for protection. This problem will probably get even worse, as we move more and more to electronic delivery. In fact, all these lower court cases have dealt with paper products, not an electronic environment. The potential is far greater with the electronic distribution.

MR. ONSRUD: I think my interpretation of the law prior to Feist was there should be no sweat-of-the-brow protection. There was quite a bit of disparity among the various circuits in the United States and it wasn't clear cut, but most appellate court cases appeared to comport with Feist. My reading of the cases is that it comes down to the way you view the data. I do not agree that sweat-of-the-brow rights were generally accepted before Feist, at least not as clear-cut law.

MR. MASSANT: Right, there was some inconsistency and essentially there was a condition of uncertainty, which is not ideal.

MS. ADLER: I want to go back to something that Eric just said, building on what Harlan said, which is fundamentally using the term that you used.

We have two very different paradigms. We have the market paradigm and the research, scientific community paradigm. I believe they are fundamentally completely at odds with each other, given what is being described. I hear Mr. DeGiusti saying there is no right to the data. We talk about that in the context of transformative use, which is fundamental to how the academic and scientific community does its business, which is completely at odds with the kind of proposal that is being made on behalf of the market forces. How one resolves those two, I don't know. Even hearing you say that the proposed legislation is not a right in data, in essence, from our perspective, that is exactly the same, as if you are saying there is no ability to use the information downstream.

As you recall, in some of the negotiations, the notion of extracting one piece of information from a database did not trigger a liability, but two could. That is two pieces of factual information could be considered “potentially substantial.” In terms of the database, that completely counters the academic paradigm that Harlan was describing. I believe that is where we are struggling.

MR. MASSANT: I agree that it is a difficult issue. Standing back, what we are faced with is, for instance, someone has invested in creating a database. In the academic market, scholars have the choice of buying the database or doing their own thing. So, to look at the private sector innovator who created this database and say, “now that you have done this and you have provided me with a choice, I am going to take your database and I am going to do what I want with it, even if it causes you harm,” that is just so inherently unfair, that I just don't see—

MS. ADLER: Let me just answer that. How about if it does harm in the marketplace? What if an academic institution licenses a database, takes a few pieces of information from a database, combines it with four others, creates new knowledge, and it is done only in the context of scientific research and education; it has nothing to do with the commercial marketplace. That is now the way the bill reads. But we are not harming your market, because it is the academic market.

MR. MASSANT: There is no problem if it doesn't harm the market. Bear in mind, there are a number of publishers who create databases exclusively for the academic market. For example, for the Congressional Information Service, which is one of the Reed Elsevier companies, 95 percent of the market is the academic market. If it is used for scholarly academic purposes, if it is released into that market, it is going to eliminate the entire market for that publication.

MS. ADLER: But it is not a competing commercial product. It is a product to extend and advance knowledge of a specific discipline. It has nothing to do with commercial activity.

MR. MASSANT: If it causes harm, it is a problem because ultimately that product will go out of business, and then where will the academic community be? Then they will no longer be presented with the option. They will have to go back and do it on their own. They are better off with the option.

MR. ONSRUD: So, what you are saying is that if this person publishes with Reed Elsevier, it wouldn't be a problem, but if they published it with some other publisher there would be a problem in that particular instance?

MR. MASSANT: I am not clear.

MR. ONSRUD: You wouldn't be harming Reed Elsevier, in the sense that you published with Reed Elsevier Publications. But if you went and published it with one of their competitors, then that would be potential harm to Reed Elsevier. So, in essence, that means that you would go with Reed Elsevier.

MR. MASSANT: That would be your own thing. We are talking about databases here.

MR. ONSRUD: Again, what is a database? If a university professor is not looking now at these contracts or copyrights and signing away his or her intellectual rights, why should we expect them to care about intellectual ownership of data sets?

Scientists generally don't care. Once their research is complete, they go on to the next topic. We are men and women of science, we communicate with each other, we want to share this information. Let's help each other in whatever way we can. So, if the licenses are too much, they will essentially go and work with a group of people who are of like mind. That is to say, they will just have to find a way to make an end run around the publishers.

MR. MASSANT: I think we started off talking about government information. Clearly, all the data sets that are funded by the government should be available to the public to do whatever they want with it. I think it is where someone has invested in creating a database, then there should be some concern and consideration for protecting the market for that product. It would seem to me that a lot of the science and technology data type of information is government funded. I don't think the private sector goes out and creates scientific and technology databases, as such, not in the context of the data sets we are talking about, because there isn't typically a market for that.

MR. EISGRAU: I have a couple of questions along those same lines. Could you tell me what you mean by taking the database to harm the market?

MR. MASSANT: As I said earlier, as far as using a database, we create databases for people to use them. If someone takes a database, uses that database, integrates it with another database, works with the data to come up with whatever they are trying to do with the data, that is fine. Where we would have a problem is if they then distributed that database, the ultimate database that they have integrated that first database into, and distributed it.

MR. EISGRAU: Are you talking about integrating the entire database, are you talking about most of it, half of it, two pieces of information out of a million?

MR. MASSANT: Integrate it enough so that, when it is redistributed, it displaces the sales of the first database. In some cases it may be that a database is created exclusively for an academic market. If that second database is then redistributed in such a way that 50 percent of it causes the academic market to feel that they don't need that first database any more because they now have this bigger database into which that smaller database has been integrated, that is a problem. Obviously, that is going to eliminate the first product.

MR. EISGRAU: So, it is an elimination or substantial likelihood of elimination of that product.

MR. MASSANT: Well, harm. I know that is an issue that the proponents and opponents had some problem getting our heads around. The extent of the harm, I think, is still an open question.

MR. EISGRAU: It is not a question of policy. I am asking Reed Elsevier, what does harm specifically mean to Reed Elsevier?

MR. MASSANT: Well, something that would cut into the sales of that product.

MR. EISGRAU: Sales.

MR. MASSANT: Right.

MS. SCHIRO-ZAVELA: On that question, I guess this is a really simple question, but I just want to understand. If I go and buy a copy of a book, it is not a copyright infringement for me to hand it to my colleague to read when I am done. If I had a database and then somebody else looks at my copy of the database, have I harmed your market?

MR. MASSANT: That is a first-sale issue. I think having someone look at the database you have is different from distributing it to others. If you are making a copy of it, that is a different issue. If you are looking at one version, you could have 10 people standing around a computer looking at the database. I don't see; that as being a problem. Once you make a copy of it and send it someplace else, that is where the problem lies.

MS. SCHIRO-ZAVELA: Maybe I am missing something. If you take a database right out of a journal and you put a routing slip on it to 20 people in a department, it seems like it would be the same sort of routing with 20 people in your department with a database, that would be harmful to the marketplace.

MR. MASSANT: A lot of databases are sold into institutions and there could be multiple users. It would be inconvenient for you to be putting a routing slip on a paper copy and then distribute it to all the, for instance, universities in the state or whatever it may be. There is an issue where the paper and the print differs from the electronic in that sense.

MR. ONSRUD: One of the issues that was raised is, could you have the right that you just described, or would you have to buy the right? In other words, if the university obtains a license for three users of a particular database, could there be a right, as long as the technology supports it, that limits access to that database to three people in the library, or three people at a time, wherever they happen to be, could essentially access those three versions or three copies at a distance, as long as it was in the library while you were essentially wherever you happened to be? Now, normally that is done through licensing, not by right, but if you do that by right, it would be closer to the current paradigm.

PARTICIPANT: Suppose that you take a database, modify it, transform it, create a product that would advance knowledge for scientists. So, it would be beneficial to society that way. Obviously, it might create some harm to the market. Is the obligation of scientists, then, who created this to say, “Here, I have done this; would you like to buy it from me or would you like to have it?” What is your obligation with respect to promulgating this? We get the same sort of thing that some people claim has happened to some extent with software, where companies will prevent innovation to protect their share of the market.

MR. MASSANT: Certainly if someone has a good idea for using our data, we would welcome them coming to talk with us about it. As far as continuing to use—getting back to what I said—if you use those data in such a way that it is going to cause harm to the original database, that is going to be a problem.

PARTICIPANT: The problem is, is it going to create a problem where you have new knowledge? That is at least one of the questions.

MR. ONSRUD: What you are suggesting is that as long as it significantly advances science, whatever the use of this database is, then it might cause minor harm and it would be the burden of the company to show that it causes significant harm and it did not significantly advance science.

MR. MASSANT: I agree with that approach. As I said, there is some kind of fair-use language in the legislation. Certainly, we don't want to prevent, obviously, the advance of science. In fact, we feel our products help contribute to it. There is a limit to it and I think at some point there has to be a balance, and ultimately that may have to be interpreted in the courts, the way fair use has been, as to whether we are siding on the side of too much damage versus what the legitimate uses are for the advancement of science.

MR. EISGRAU: The premise of one model would be that protection does not apply unless the owner of the information can come in and bear the burden of showing that science wasn't advanced and significant harm occurred. In fair use, by definition, the user is on a defensive posture. Could you comment on the potential utility of the model for database protection?

MR. MASSANT: I think in database protection the model is somewhat flipped, in which the burden is on the publisher to show that there was harm. I think that is the way it has been structured, more or less, as opposed to being a defense to determine whether use is fair or not. So, those uses could be made, and then we would have to show that there was harm. In fact, continuing on the same thing, in the database protection, as far as not-for-profits are concerned, they were exempt from any kind of criminal liability and financial damages were remitted in cases where there was innocent infringement by a not-for-profit.

I think there was a question earlier about liability. I think the liability was reduced significantly because there was a concern in the not-for-profit sector for being uncertain about how liable they were. So, the effort was to reduce that significantly, and that was certainly evident in the last version of the legislation.

MS. ADLER: Recognizing that this isn't the last version of the legislation, what I was confused about was your response. This version of the legislation says that any person who extracts or uses in commerce all or a substantial part of a database, measured qualitatively or quantitatively, and so on, then the person shall be liable. So, the person is liable. So, the burden, it strikes me, is on the individual, or on the institution and not on the publisher. That is why I am confused about your response.

MR. MASSANT: If somebody uses the data, it is up to the publisher then to determine whether there has been some damage, some harm caused, and then to take action against them. The burden then is to prove that there has been harm done. I am not a copyright lawyer; I am not a lawyer. It seems to me that that puts the burden on the private sector, to show that there has been harm caused.

MR. ONSRUD: There are lawsuits now against university professors making multiple copies of things that many of us thought would have fallen under fair use. So, in one respect, legislation like this would, with resources that the commercial sector has, ensure that these cases would be brought. Of course, customary practice is not part of this database legislation, the way it is proposed.

MR. MASSANT: Isn't it correct that in fair use the burden is on the user to show that they had made fair use?

MR. ONSRUD: Well, there were lawsuits brought against some of the individual faculty at some universities who thought that they were making fair use of the materials. They lost the cases and those precedents had ramifications at the universities as far as the way things are now done. They are operating in a very cumbersome fashion.

DR. MC DOWELL: Essentially under compensating for harm, you might want to consider taking a look at real property law. There are established principles and constitutional principles that the public may take private property for public purpose. The only question is how much compensation. What is new in recent years is that that philosophy is being applied more and more to regulatory actions, such as zoning, wetlands regulation, and so on.

The question becomes more and more, Where is the tipping point? How far can you go in a regulatory fashion before you essentially trigger confiscation of property and must pay for those partial rights in property that you have taken by regulation? I don't know if there is a commonality of thought there, but it appeared to me to have some common threads.

MR. EISGAU: The Supreme Court has repeatedly rejected it, although it is a useful analogy. The underpinnings of the intellectual property system in America have certainly manifested themselves in copyright. Intellectual property is fundamentally different than the policies and the laws underlying real property, protection. The Supreme Court has repeatedly stated that information is a public good, and the purpose of the copyright law is to provide an incentive for people to create stuff so that public good can become available down the road.

Recognizing here that we are talking about potentially providing more protection for stuff that doesn't meet the constitutional standard for copyright, the question is not how do we analogize to the real property model, but rather, how do we stay true to the Constitution with regard to intellectual property policy and, in effect, maximize the creation of information and the public good that comes from it. So, there is some utility to that comparative model of real property, but recognize that for very good policy reasons, it doesn't have a foundation in intellectual property law. Intellectual property, despite assertions by people to the contrary, has never been treated that way.

MR. DE GIUSTI: I must say, the database protection is of a short duration, unlike copyright, which offers protection for life plus 70 years now, or 90 some years, and 120 or whichever is shorter for works for hire. As such, the proposed legislation is of much shorter duration.

MR. ONSRUD: Any last comments here?

DR. SCHOOLMAN: One issue that seems to have gotten short shrift in this discussion has to do with preservation, maintenance, and duration of a scholarly record. The duration of 15 years is totally incompatible with the creation and the maintenance of a scholarly record. I think in some form it has to be, in the future, a critical element of the scientific research picture.

I have ignored the investment issue because that already is an important issue, which is going to be presumably thrown into the bargaining or the licensing that addresses it. My main concerns at the moment are archiving, preserving, and maintaining the scholarly record. In some way, the exemption of those who were engaged in that activity from extraction and publication of the extracted material should be prohibited under the current proposals for the database protection. That issue needs to be worked out in some form or another.

MS. ADLER: I don't disagree that this is an issue that needs to be addressed. I am trying to think of this issue in the context of what we have been asked. I think it is a separate issue. I agree with you it is a crucial issue. What I am not clear on is how one would draw the line.

Speaking for the Association of Research Libraries' institutions, the preservation and archiving is not the only function we play. It is robust access as well, which is the other side of the coin. That then would trigger the kinds of discussions that we have had this morning. I am just trying to sort my way through your question in that regard. We did in the Senate negotiations specifically ask for library exemptions for preservation and the like.

DR. SCHOOLMAN: I think you stated my point, in that if we are going to make exceptions and not simply eliminate the full database protection concept, and we are going to create a database protection law—which is, I think, likely—then we are going to have to try to create a series of exemptions from that protection. One exemption that I think needs to be considered among the many that you have already considered is archiving and preserving data, which hasn't been considered or at least not been given enough attention, in my opinion.

MS. ADLER: I understand; I absolutely agree with you on that.

MR. FRAZIER: One quick reference to the apparently perpetual nature of the new rights that are being created for databases: These rights appear to have no time limits at all, because every time the database is renewed, you can continually extend the protection of the entire content of the work. This would really overthrow the entire history of American copyright law, which was born in the limitation on the terms, of copyright. It is absolutely fundamental to our intellectual property system. It is clear that this is aimed not at competition, but at users of information, in an attempt to create perpetual rights in intellectual property.

MR. ONSRUD: One question about the proposed law. Are publishers, in order to gain copyright for a particular database, required to file the data? If a producer of a database is required to deposit that database and archive it in order to gain protection for it, and then, to ensure that it is in the archive 15 years later so that people will have access to it, that would ensure that indeed every version of this database was protected and was essentially in the archive.

MR. MASSANT: That is an important point, because the protection applies only to the new data in which there has been a significant investment, not to the material that is older than 15 years in the database. Of course, the statutory protection does end after 15 years. So, in fact, the idea of depositing databases so that people could be absolutely sure—if we did get into an issue of how is a user going to be able to tell what is new and what is old—that is very difficult to do, so one of the alternatives that has been considered was that we would deposit our databases and, at the end of 15 years, they would be fair game.

MS. ADLER: I think that is a critically important part. That was the very issue that the FTC saw, as many of us did, as perpetual protection for the database. There was a provision—it was not as strong as we would hope, but it was moving in the right direction—included in the Senate negotiations for deposit at the Copyright Office.

As the legislation was finally drafted, there were still, including at the FTC, very strong concerns that, indeed, protection could be perpetual. Under any kind of regime discussed, I think that would have to be clarified right up front, for the concerns and reasons that Ken discussed.

MR. ONSRUD: There is no guarantee that after 15 years there won't be legislation for another 15 years and another.

MS. ADLER: Déjà vu all over again.

MR. ONSRUD: I have gained a lot of information here today. In my own mind, I don't know if this has clarified the issue more or if I am more confused, but it has certainly been interesting.