Paper presented to the ICSU Press Workshop, Keble College, Oxford, UK, 31 March to 2 April 1998
Electronic publishing is not going to replace the need for eyeball to eyeball discussion of tough issues, nor will it replace the role of beer in facilitating the birth of new ideas. With those worries aside, we can pursue our work here without fear of destroying the climate in which science flourishes.
In publishing science there are some immutable objectives and guiding principles. Within these boundary conditions the comfortable, and up to now apparently orderly, business of publishing is an evolving non-linear system undergoing destablization.
Good - this approach to chaos may be necessary for new ideas and different methods to flourish. With publications loose in bits and bytes we open up all sorts of new opportunities to meet the needs of science.
My understanding is that we are here in Oxford to take a quantitative look at the costs and benefits of electronic publication in science. I wonder how many different views there are in this room of what defines success in publishing science. Can we agree on the elements that are essential to maintain an adequate science publishing system and specify some of the benefits electronics can bring to the system? If so perhaps we can make real progress. Science publishing exists to communicate science effectively between and among individual scientists everywhere - across borders, oceans, and cultures and anytime - now and into future generations. In my view none of the existing institutional participants in this system including libraries and publishers have a preordained role as we continue to use more and more new technology in communicating science. Costs must be based on value and benefits. The scientific communication system must not be hamstrung by the needs of institutions to perpetuate themselves.
Many individual scientists have a vision, which I exaggerate only slightly, of a communication system that provides precisely what one is looking for, instantaneously, at no cost, and in a choice of formats including that which will serve the particular need one may have at the moment, whether that need be browsing for inspiration, references for the paper in preparation, a mathematical model to be tested, data on the magnetic field in interplanetary space, or a print copy of a paper to read on the way home tonight. All of that is within our grasp today. But -
One of my early lessons in business came at a printing plant in the form of a sign, presumably for the visiting customers. That sign said: price, time, quality - pick any two. Although that maxim has been confirmed time and time again part of me still rejects the 30 year old lesson as a defeatist attitude. I am prone to believe that there is always a better way. On the other hand, I muse, perhaps this is just another restatement of the three laws of thermodynamics, which, at least for now, I still hold to be inviolable.
The restatement of the three laws that I like best is - You can't win; you can't break even; and you can't get out of the game. I'd say that pretty well describes where we are in electronic publishing today. Our exploration of this jungle is costing a lot of money and there is no alternative.
The vision of the scientist is before us and we are so far from it that we needn't worry about the thermodynamic limits. At the AGU we are putting a lot of energy into making as much of that vision a reality as possible. But I don't know any better than the myriad of scientists I talk to how to get from here to there. The only difference is that I am certain I don't know, and I am prepared to argue here that the way in unknowable.
Attempts to define a single path will prove fruitless. It is patently stupid in an environment that is evolving as rapidly as our electromagnetic environment is, to look for solutions that will satisfy for 100 years or perhaps even for as much as 5 or 10 years. There have been efforts by large organizations to subjugate the author and user communities through proprietary hardware and distribution schemes that would give these companies a significant advantage. These dreams are being challenged by the tough spirit of openness and individuality that pervades the electronic frontier.
There will be continuing change, forget transition, and don't look for stability. Scientists, publishers and librarians all seem to suffer delirium when it comes to electronic publishing. And outside our world no one gives a damn. We are swept up in a maelstrom of constricting and conflicting business and societal imperatives. As their balance shifts our boundary conditions change. Legislation, treaties and corporate policy can suddenly, and inadvertently, alter the course of scientific communication and perhaps that of the entire scientific enterprise.
This is not a manageable environment; it is one we must learn to live at peace with.
There are many complications inherent in the electronic environment, but they are only a part of the forces that will be driving change, in fact mandating continuous change over the foreseeable future. The technology will evolve, and with each generation new services will have to be offered. Classes of institutions may find themselves unable or unwilling to assume the burdens of knowledge handling that they have in the past. An ever more broadly distributed customer base will both strengthen the system of scholarly communication and introduce further uncertainty. The complexity of our knowledge-handling job will multiply enormously. And finally we must fit into an uncontrollable environment.
Let's start with the technology; how fast is it going to change, how different can it be, and what difference is that going to make to scholarly communication. We have had journals in printed form for something like 300 years. Microfiche and film was the answer of the 1950s. Few would think of using it today, and what a headache it is going to be to convert that microfiche less than 40 years later. Now, convert it to what, CD-ROM? How long does that technology have left to run? Not long, I'll bet.
We are all familiar with the exponential growth of computing power per dollar as a function of time. What is going to make that stop? Band width has likewise grown. These are incremental changes, but combined they provide for continuing breakthroughs in communication. On top of them, there will be new breakthrough that are going to shake us up as much as the computer has. Try this - How long do you think it is going to be before we are able to access the brain directly and what will that do to change publishing is science?
Whatever new forms scientific publications take next you can bet they won't last long. Certainly not 300 years and I would bet on something closer to 3 than 30.
The communication channels of science are diverse, and in many cases, we can't be sure of what we need until we actually see it. Currently most indexes provide only hints as to where to look beyond the formal literature. A large fraction of our needs lie in organized and disorganized data repositories, gray literature, scientific meetings, and chitchat. In the electronic environment the boundaries between the "formal literature" and other communication can be blurred or erased. Active scientists need to integrate them all and because we can help move in that direction we are expected to. Paul Ginsparg's enormous preprint archive at Los Alamos is one individual's idea of a start on that road.
Not only is it going to be difficult to satisfy the demands of users for the kind of access they think they should have to the storehouse of knowledge, but also the contents of that storehouse are going to get more complex as technology permits and science makes use of the technology. We already have to deal with audio and video in the formal literature. We must insert and maintain links to other materials in the record and provide the ability to interact with equations, figures and tables. It is impossible to say where we are going from here except to say that our future is likely to be more complex and more expensive. The control of the scholarly communication system is passing to the hands of the user - where it should be. Others must lead, get out of the way, or be run over.
Just as there is no single paradigm for the financing of scientific publication today, there will be none in the electronic environment. To the contrary, more flexibility is offered and information providers who are now strictly on a subscription base are likely to be forced to provide per copy, per view, and maybe per minute services. Up front charges and advertising will also be part of the mix. You can be sure it will be a mix; it will be tuned to the customer; and perhaps to the individual user/scientist.
Perhaps one of the most unsettling thing for some publishers will be the question of who is the customer. Many publishers have viewed the library as the customer. This stable, high-paying customer (as we know him) may well be on the way down the drain as individuals gain direct control of their resources and expenditures. That thought should raise really scary scenarios for subscription based publications in highly specialized small disciplines. The few practitioners in the field may have to individually support their communication system and the archive that has heretofore been the responsibility of education and research institutions. Publishers may have to live on the judgment of individuals as to the quality of each paper they publish. Perhaps some will suffer an overdue demise.
Let me talk about the archive a bit. It is the storehouse of the body of knowledge and absolutely critical to the advance of science. The fundamental question here is where can the ultimate responsibility for safeguarding and making available the record of knowledge reside?
Technology has completely changed the problems of storage and access and as I said earlier further change is the only certainty.
Thirty years ago the world's academic libraries and the institutions that stood behind them had this trust and few gave the problem a second thought. There was substantial redundancy and funding seemed adequate to maintain the situation - at least in science. This was at the end of the era in which Maxwell's law was developed. Your,more recent, is reported to have said "there are a thousand libraries that will buy any scientific journal". It appears that today libraries are focussing more on providing access to information than on storing it. Only a few people seem to be asking whether there will be anything to access. There are many journals distributing a few hundred or less copies worldwide. As the number of subscriptions continues to decrease these will be on the endangered list. In the electronic environment every document will be endangered from birth. Will there be anything to access if we do not begin to deal with the storehouse soon?
The list of potential participants in maintaining the body of knowledge includes: government entities; not-for-profit educational and research institutions; corporations with a stake in a particular segment of knowledge; corporations organized to maintain a portion of the knowledge base on a for profit basis; scholarly societies and individuals scientists themselves.
Most of these have fatal flaws that rule out their providing a dependable archive into the future. To solve these problems the storehouse is likely to be a vast web of specialized sites that will provide redundancy in the way libraries do for printed materials today. The sites will have to be specialized because of the inevitable plethora of technical requirements to keep different parts of the literature alive electronically. Common standards can only be at the most basic levels because of the demands of the different disciplines. I expect that the continuing strength of this web will be based on the involvement of individual scientists who have a real and immediate stake in the outcome. The cost of maintaining the archive and access to it will be a major component of the overall cost of electronic publishing.
Governments, including, if not exemplified by ours in the US, are notoriously unreliable as trustees of the body of knowledge, both in terms of providing the resource and in terms of exercising the judgement required. Benevolent rulers are a rare commodity today, therefore, at best, government decisions are dictated by majority will. Intellectual and cultural matters rarely fare well. In this milieu decisions are often made that support political or religious ends that are antithetical to science. Governments through the ages have been known to pare and shape the knowledge base to their own ends. Security classification, which is a real hindrance in advancing some sciences, is perhaps the most benign of these kind of intrusions.
The incentives in the corporate world militate against any undertaking that does not contribute to profits. Should our storehouse of knowledge be at the mercy of corporate bean counters? One group that might seem logical for providing some of the archiving function is the information providers. Most of them, however, are currently just intermediaries, or perhaps even parasites, that rely on the resources and markets of others. Their motivation to retain archives that are not producing income is difficult to see.
Perhaps academic libraries with the support of their home institutions will develop income generating operations supplying information so as to fund the conservation of segments of the knowledge base that are of special interest to them. At some level this library role must continue as a natural adjunct to serving the information needs of their internal research communities. But, at what level? To what extent will the library become merely a conservator of artifacts and an information intermediary?
Who has a real interest in conserving the body of knowledge? It seems to me that only those who care about the field have a consuming interest in the body of knowledge in a particular area and the expertise to maintain it. And, even under their trusteeship there are risks; for example risks that current biases and fads will dictate what and how things are preserved.
To find the ultimate trustee we must look at where the incentive is. Who will feel the pain if the job is not done well. Those are the only people likely to pay for it in a squeeze. The person with the strongest interest in maintaining the body of knowledge in physics is a physicist who has contributed to it and whose life revolves around it. The further one gets from dependence on that body of knowledge the less interest one has in sustaining it. In the final analysis, it is the community of individual physicists who are likely to be the best trustee for physics. And who, at the same time, might be happy to let much of psychology, economics or even geology slide away into oblivion.
Individual scientists and their voluntary associations can work with numerous other organizations in different disciplines and of different kinds to sustain and protect the body of knowledge. They are the only ones who can set the criteria and standards for their disciplines and who will assure these these criteria are being met by themselves continuously monitoring and filling in the gaps. If this sounds a bit filmy or flimsy to you, then you are getting the picture.
While I do not think we can get a clear picture of even who the players will be in the next few generations of the knowledge handling and scientific communication system, we can say a few things with relative confidence. The individual users and creators will play a more visible and larger role than they do now. New and old players will move in and out of the system as their ability to meet changing user expectations waxes and wanes. Change will be driven by user expectations which will stay ahead of the capacity to provide the desired service economically. We will be writing off hardware and software in less than 5 years and most of our best ideas in far less than 10. Little providers will be very important to the extent that they represent specialized groups of users with special needs. The giant publishers of today will go the way of the dinosaurs unless they learn that libraries are not the end user of their product. We are entering a world where value rather than library inertia will sell the product. The harsh light of an open electronic market will reveal value added as well as lack of it. That value is, at the end of the day, all that is offered to the individual creator and user.
In the electronic environment, the end user is going to have to be treated as an individual. Every individual out there is a potential predator, harasser or parasite. Each of them has the tools to enter and compete in most aspects of the scholarly communication enterprise. Only when every scientist who contributes to the body of knowledge and everyone who uses the storehouse have really good feelings about their interactions will we be near equilibrium.
My view of the future may appear to be chaotic. It will provide little or no comfort to a linear thinker trying to build a business in information dissemination, nor any haven for the risk averse. The pathways from creator to user will be multiplied many fold. The challenge to those of us who are seriously interested in serving the science and not just in extracting money from the system, will be to encourage experiment that enhances value to the user and creator whilst meeting the standards of the scientific community for the knowledge base.
Any meaningful discussion of the costs of electronic publication in science must be based on a definition of what is expected in terms of values and benefits. Let me suggest core values that can serve as an anchor that will allow us to discuss all kinds of current ideas for the directions that electronic publishing may take, and leave open the possibility of developing the science publishing system in entirely unpredictable ways and adjusting our economic models as we go along.
For your consideration I suggest the following with no claim to originality or pride of authorship:
1. Certification: there must be a process by which a contribution to the literature is filtered and given a seal of approval that says that this is a new contribution and significant enough to be worthy of preserving as a part of the body of knowledge.
2. Dissemination: the scientific literature should be available to scientists worldwide.
3. Authentication: assurance that the record of ideas, discovery and hypotheses tested is not altered.
4. Continued access to the record: the record must be preserved and be accessible to future generations of scientists into an indeterminate future.
Of these four values only authentication is a black or white issue. Within the others there are not yet absolute, or even well defined, standards that define the acceptable level of performance.
Quality standards in all these areas should be established by the scientific community. Standards will help fine tune models of the communication process and cost projections. Without agreement on what constitutes responsible and acceptable scientific publication at the basic level of the core values there is no way that progress can be made in developing meaningful estimates of costs.