How to Think about Strong Collaboration among Professionals

Larry Sanger

Keynote delivered (omitting section 4) at the Handelsblatt IT Congress, Bonn, Germany, Jan. 30, 2007.

1. Introduction: the rising tide of strong collaboration among professionals.

Nearly everybody knows that Wikipedia is a giant free Internet encyclopedia created by a worldwide collaboration. What if there were similar, Wikipedia-like global collaborations for every profession and global industry? Our professions and industries worldwide could use the same sort of collaborative techniques to create new kinds of resources that would be enormously valuable. Imagine what people in your industry, or your profession, could do if very many of them were ready to get together to collaborate online, Wikipedia-style.

Now, if I put it that simply, my suggestion is too vague to be useful. So what I�m going to do in the rest of my talk is to clarify this suggestion.

What I want to talk about now is how we might organize professionals from many different companies and from many different countries to collaborate on information resources that everyone can use, presumably for free. For example, imagine an international federation of journalists getting together to exchange information and summarize ongoing news stories, so that they (and everyone else) could find the latest information about a developing story by looking in one place, rather than having to read many different articles.

To take another example, you can easily imagine lawyers and law students working together on a system that allows them all to summarize cases, to "mark up" legal arguments in ways lawyers find useful, to place cases into different categories for easy searching, to integrate all this information with law journals, and so forth.

Now, of course I know that there are very many proprietary information resources that are ultramodern and extremely useful. I don�t mean to argue here that we can replace those resources by collaborating. I am saying, instead, that there are new sorts of information resources that can be created only by very large numbers of people collaborating. Many of the possibilities haven�t even been thought of yet. Before 2001, no one really knew that a Wikipedia--an information resource with over 1.5 million articles in the English language alone--was even possible. But it is. And as it turns out, Wikipedia isn�t just a big encyclopedia. The thing is actually so enormous that it isn�t really an encyclopedia in any ordinary sense. The sheer bulk of it makes it a brand new sort of resource.

Still, you might be skeptical about Wikipedia. You might be skeptical whether such open, collaborative systems in general can be made to work. For all the visionary talk I like to indulge in, I am naturally a skeptic as well, and I think people will have to fail in all sorts of interesting ways before they create systems that work well. We are very, very early in the game. Clearly, we have much to learn. Most of the possibilities that I (and many others) can imagine have not even been tried.

What�s really exciting to me is that a growing number of professionals are interested in exploring these possibilities. In the last two or three years I have been approached by quite a few different people and groups, from many different fields, who are interested in starting broad-based, open collaborations for different professional communities and for different purposes.

One of the best examples I can give you is the Community Patent Review project, which is still in the planning stages. New York University Law Professor Beth Noveck sees a serious problem in the United States patent system, which takes far too long to decide whether to grant a patent, and then frequently grants patents for things that are, to engineers, completely obvious. The solution, Noveck suggests, is to recruit the very people who use the patent registration system to give feedback on patent applications. By opening up the process of patent review to collaboration by a potentially huge body of engineers and inventors, the delays and inaccuracies might be greatly reduced.

I can give other examples:

• I was approached by a U.S. government food safety expert who wanted to start a network of food safety professionals, who would maintain together, wiki style, a single database about contamination outbreaks.
• A few years ago, I spoke to some business people who wanted to create a wiki or a blog--they weren�t sure which, maybe a combination of the two--that would help the development of technological innovation in the state of Ohio.
• Last year, I helped to plan a wiki called the Encyclopedia of Earth that offers free information about environmental issues. Its editors thought, correctly I believe, that the community of experts studying environmental issues could be recruited to build a really reliable body of free information for the world, and to work on the articles collaboratively.
• The National Institutes of Health, a U.S. government research body, is holding a "wiki fair" in a month. I�ll be speaking to them, and I can tell you that they�re quite interested in aggregating medical research using wikis.

Actually, these are only a few examples from my own limited personal experience. If I wanted to bore you to death I could give you a few dozen more examples of professional collaborative projects I�ve heard of, or that have been under discussion. The fact of the matter is that there is a growing movement to bring together the efforts of professionals, who want to do for their specialized interests what Wikipedia has done for information in general.

Here then is what I want to do in the rest of this talk. First, I want to explain what "strong collaboration" is. I�ll illustrate the concept by reference to two projects I�ve helped to start, Wikipedia and the Citizendium, or the Citizens� Compendium. Then I will give you several ideas for professional collaborative projects which I don�t think anyone has tried to start yet, but which think perhaps somebody ought to try to start. Next--if I have time--I will offer a few principles about how and how not to start strongly collaborative projects. Finally, I will speculate about what a future with massive, thriving professional collaborations might look like.

2. Old-fashioned collaboration and "strong" collaboration.

If there is one thing, more than any other, that makes Wikipedia really productive, it is the fact that it is strongly collaborative. So let me start by explaining this distinction between old-fashioned collaboration and what I call "strong" collaboration.

The prototype of old-fashioned collaboration is a report written by committee. This often involves one person assigning parts of the report to different committee members. The editor stitches the parts together, sends the result back to the committee, receives some changes, and then issues the final version. The result is often so boring as to be unreadable. An interesting thing about old-fashioned collaboration is that it�s not very collaborative. One person writes one section, and in many cases no one else even changes a single word. There is multiple authorship, but there isn�t shared, joint editorship. People don�t really edit each others� work.

Now, a more robust sort of collaboration takes place between two or more people who are writing, or creating art or music, side-by-side, each responding to the other in real time. They are not just contributors; they are also each editors of the work. There is definite collaboration going on, and it�s more robust than committee-style collaboration, but this too is old-fashioned.

Strong collaboration--which is made possible on a wide scale by the Internet--goes one step further. Not only are there multiple authors, and not only are those people each others� editors, but there is no set group of people who are the authors and editors of the work. Some authors can drop out, and other authors join. In five years, you might not know any of the people maintaining a strongly collaborative work that you started. (It�s a little like the evolution of folk tunes.)

On today�s diverse and creative World Wide Web, there are many good examples of strong collaborations, but wikis in general and Wikipedia in particular are perhaps the best examples. On Wikipedia, articles are written by whoever comes to a Web page and decides to press the "edit" button. Wikipedia features multiple authorship and joint editorship, because many people work on it together and they edit every part of each others� work. But each article also has an ever-changing, rotating group of maintainers.

I think its policy of strong collaboration is a large part of the reason that Wikipedia is so huge. Virtually anyone can come to the website and work on articles. And, for whatever reason, a lot of people do want to work on an encyclopedia. With Google spidering the website regularly and sending new contributors as it grew larger, it was only a matter of time before the website became huge.

But you might think that Wikipedia�s strong collaboration, which partly explains its enormous productivity, also explains its infamous unreliability. I don�t really agree with this explanation myself, but let me explain the objection a little more. Since anyone can participate, all sorts of ignorant and biased people show up and degrade the quality of articles. Moreover, with the bar to participation set so low, many people who have higher standards are not interested in participating. So, on this view, Wikipedia�s quality will never rise very high.

But actually, Wikipedia�s quality is remarkably good for a system that is as open as it is. Indeed, Wikipedia demonstrates that openness is compatible with fairly good quality. One of my main objections to Wikipedia is not that it is too open, or too collaborative, but simply that the system makes no special roles for experts. I think it is possible to have a very open and strongly collaborative system that has special roles for people who know a lot about a field.

In fact, that�s what I�m trying to accomplish with a new project that I first announced last September in Berlin, called the Citizendium. It�s very similar to Wikipedia, but we require that people use their own real names--no anonymous contribution--and we have expert editors who work on the wiki right alongside everyone else. These editors can approve articles, and they can make decisions about the content of articles, but they have to work with everyone else.

The Citizendium is now in its pilot project phase, and I can report that it�s going quite well. We�ve produced a respectable number of articles in a short time, and our articles tend to be quite good. We have shown, to my satisfaction anyway, that this sort of expert-guided strong collaboration can work.

If the Citizendium can work as well as Wikipedia--only time will tell that for certain--that would indicate that professions and industries can create strong collaborations themselves. So next, let me give you several ideas for how, in my opinion, professions and industries might create great information resources through strong collaboration among professionals.

3. Some ideas for how professionals can create strong collaborations.

Several questions will help us focus our thoughts on what professional collaborations are worth pursuing.

Here�s the first question: "In your industry, whether it�s finance, or manufacturing of some sort, or information technology, or whatever, what kinds of information are of interest to everyone in the industry?" To a certain extent we can make some generalizations here, regardless of industry. Everyone in all industries is interested in best business practices, the latest product information, books and articles of interest, what the industry leaders (both companies and people) are doing and saying, and so forth. There are many classes of information that are of interest to everyone.

Another relevant question is this: "What sorts of information do people collect or review, which other people collect and review in pretty much the same way?" Things like bibliographies, summaries of important articles, and general introductions to a topic are written and compiled over and over again by different people for their own personal (or company) use. It can be helpful to have these people work together, then, rather than endlessly duplicating effort.

My final and perhaps the most important question is: "For each of those types of information, what sort of information resource can be created only using methods of strong collaboration?" In other words, I�m asking: what sort of information resources can people in your industry create only if very many of them work together as parts of a huge, ever-changing, rotating group of information maintainers?

So I think there are three criteria that can be used to identify particularly good candidate projects for strong collaboration. First, the project should aim at compiling information that is of interest to a lot of people. Second, it should be the sort of information that very many people regularly collect or review, duplicating each others� effort. Third, an exhaustive repository of that information can be created only by methods of strong collaboration. If all three criteria are met, you�ve got an excellent candidate for strong collaboration.

A wiki that is used to compile best business practices for an entire industry would be an excellent example. Not coincidentally, the very first wiki, called WikiWikiWeb, was itself a wiki about best software practices. Wikis are actually a pretty good medium for writing "how to"-type material. Probably, only a large group of collaborators can really maintain an excellent body of information about best practices, because best practices themselves change constantly, in most fields--and the information is in many cases spread out among very many different people. In other words, people have all different sorts of ideas about how to do things in their fields, and because, in order to stay competitive, they�re constantly innovating, they�re also constantly thinking up new ways to do things. On the other hand, many established practices are no doubt endlessly reinvented due to ignorance of where to go find out about them. So there�s a great deal of duplicate work, in part because reliable information is hard to find.

There are many other examples. It is only through a strong collaboration that people can pool a very large number of opinions on a regular basis about what books or articles are worth reading. Here I can give you another example. There�s an Internet service for scientists--subscription-based, so, not free like Wikipedia--called Faculty of 1000. In this service, scientists recommend to each other different articles. This simply could not be achieved without the participation of scientists who score different articles, and without those scores then being aggregated. That�s a kind of strong collaboration, I would say.

There is another class of information resource that I think could not be done, at least not optimally, except by an enormous collaborative effort: a really thorough, comprehensive, and balanced review of every part of the literature of a field. In other words, a comprehensive account of the latest advances in a field in the last, say, year. It is very labor-intensive for just one person to create this sort of overview of the literature even about some very narrow topic within a given field. But, of course, together, professionals know the literature of their fields far better than any one of them knows it. Furthermore, if they are writing summaries of the latest research, their summaries will be far better if they can correct each others� mistakes. And bear in mind, also, that anyone who is a real professional in a field has to keep up with the latest advances--they all have to go over a lot of the same material, so they�re all doing the same thing. So a collaboratively-written, comprehensive summary of the literature of a field makes sense for all the reasons a strong collaboration is recommended. I would absolutely love to see such reviews of the literature about my own interests of philosophy, the Internet, and Irish traditional music. I think it is only a matter of time before people start very regularly producing literature reviews collaboratively.

So far I�ve only given you examples of some very general types of information resource that have application to many different fields. There are also types of collaborative information resources that I can imagine that would be specific to just one field or a few fields.

For example, consider journalism. Journalists collectively--but not collaboratively--create an enormous output of news, but it tends to be very repetitive. Imagine a news service that collects and unifies the news contained in newspapers, magazines, wire reports, and other sources. The service would for example let you see all the news about a certain European Union proposal, summarized in one place. I don�t mean you could read the original articles in one place--such services already exist--I mean that the actual original reporting contained in those articles would be summarized and integrated into one big document, with the source articles referenced and linked from that document.

In other words, it would be a universal digest of the news. It would be a huge job to produce this. In fact, I think it would be too large of a task to be done by any one company, if the company wanted to do a good job; there�s just too much news coming out every day. In fact, I doubt that journalists themselves would have the time to do it, in addition to all the reporting they do. I say this in part because I myself spent several months doing nothing but summarizing news about the old Y2K problem, and I can tell you that it is very hard work.

The only way to create a universal news digest, I think, if it could be done at all, would be to enlist enormous numbers of volunteers. And then, to attract the volunteers, the digest would have to be free. You couldn�t get those volunteers to work on it if it weren�t free for everyone, because very few volunteers are willing to work on behalf of a for-profit subscription-based service. But, I think there should be some journalists involved, because if there is one thing that the movement toward citizen journalism has shown, it is that few citizens are actually very good at doing journalism. Even granted that people will correct each other�s work in a strong collaboration, editors are still needed to, at the very least, set an example and clarify standards. Anyway, I think that a collaboratively-produced universal news digest is a very exciting idea, and I myself might try to organize it in a few years, if no one else gives it a good try.

I could go on and on with the examples, but I think I�ve given enough. It isn�t hard to think of such examples. It just requires a little imagination and creativity.

4. A few thoughts on how to set up strong collaborations.

At this point, I have no idea whether I�ve convinced you that strong collaboration among professionals is a good idea. But I�ll pretend for the sake of argument that you�re interested. What I want to do next is to give you a few thoughts on how to set up strong collaborations. I�ve been involved in organizing several of these projects, so I have a few ideas about how to do it.

First, what I think many people don�t often realize is that starting a collaborative project is not unlike starting a business. This means, among other things, that there is at least a potential demand for what you want to build. Also, you have to have either a "niche," something that no one has ever successfully tried, or else a better approach to something that other people are already doing. Then, when you�re actually pursuing the project, it is extremely important that you focus on the basic project, and not get side-tracked with all the other interesting ideas that are in the near vicinity. I suppose all of this is completely obvious to you business people.

But maybe the next points will not be so obvious.

If you�re going to start a strongly collaborative project, that means you actually should collaborate. Some people seem to think that it�s enough simply to install a wiki, for example, or some other content management system, and collaboration will just happen automatically. Well, not necessarily. Just because you install what is called "social software," it doesn�t follow that people will actually use your system to collaborate. Many people still don�t have the faintest clue about what strong collaboration looks like. They have to be taught not only how to use the software, but also some concepts of collaboration that are very unfamiliar.

For example, one of the central concepts it�s necessary to teach your participants is that, if they really do want other people to help develop information that they have added, then they have to leave their work unsigned. Placing your name on an article, for example, strongly discourages other people from working on it, even if the community rules say that others can work on it. Therefore, for maximum productivity, strongly collaborative content should be unsigned. Of course, this doesn�t mean that you shouldn�t get credit in a project log that tracks who has done what. It means only that the units of information presented to the end user should not be labelled with any particular collaborator�s name.

The project should be collaborative, but it should also have definite and active leadership. One thing I�ve discovered repeatedly, in the projects I�ve organized, is that people need a certain amount of structure for them to "play a game" on the Internet. They want you to tell them what the rules of the game are. They won�t simply organize themselves, usually; the participants who are happy to collaborate on information collection, most of them, won�t also be very interested in collaborating on system design. So the project leaders, to a great extent, have to set out the rules of the game. Now, it is an excellent idea to consult with your future participants about the structure of your system. You should make people feel that they are part of the system design process, and you should listen to them when they do have something to say, and be prepared to take their advice. But for the most part, most people will expect you, as project organizers, to make proposals and to lead the way. It�s only after some definite system is in place that many people will feel comfortable collaborating.

Now, in creating a strong collaboration, you might think that you�re simply creating an information-gathering system. But you are doing much more than that: you�re creating a community. A strong collaboration isn�t an impersonal "content management system" that somehow all by itself creates loads of content--it�s actually a group of people. In a strong collaboration it is people first and foremost that are your engine of production. These people, to be maximally efficient and productive, need to have rules, but not too many rules. Generally, you can entrust decisions to the judgment of well-meaning volunteers; they will usually make the right decisions, particularly if your project is open and transparent, as strongly collaborative projects must be. If people are constantly correcting each others� work through strong collaboration, it is not necessary to have complicated rules and processes. The sort of rules that are necessary are, mainly, rules about information standards and about unacceptable behavior--in other words, the acceptable moves in the game, and what the rules for good gamesmanship are. The sort of rules that can be harmful to the productivity of strong collaborations are those that define processes that can only be executed by a limited group of people--by a bureaucracy. The bureaucracy thus becomes an exception to the general collaborative rule, and becomes a bottleneck to development. Sometimes, of course, some bureaucratic oversight is inevitable. But top-down bureaucratic intervention is in direct conflict with the freedom that makes strong collaboration work. So it must be kept to a minimum.

Even if you set up a system and you define some reasonable rules, that doesn�t guarantee that people will actually use your system. It is absolutely essential that project organizers publicize the project and then communicate regularly with the participants, to motivate them to contribute. It�s also very important that organizers actually participate themselves. If organizers don�t participate, then it will look like they don�t really care about the project. They have to take the lead.

There are, finally, two specific policies that will, if you adopt them, tend to maximize participation. The first is a neutrality policy. This means that no points of view are privileged in the system. By remaining open to a wide diversity of opinion, the project then remains attractive to the largest possible group of people. The second policy is to release the information under what is called an open content license--in other words, make it free, and guarantee to contributors that it will remain free forever. This will not only maximize the value of your project to the world, simply because many more people will use it than if they have to pay to use it, it will also maximize the number of contributors. People will of course become contributors far more readily if they know they are benefitting the entire world, and not merely lining the pockets of a profit-making enterprise.

5. What might a future with massive, thriving professional collaboration look like?

I want to conclude by telling you a story, a science fiction story. What would our future look like ten years down the road, if strong collaboration among professionals were then thriving on a massive scale? It might look like this. You�ll have to forgive as I now let my imagination run wild.

"It is now the year 2017. There have been practically unimaginable amounts of information on the Internet since the early 2000s, but much of it has been of questionable quality. In recent years, however, professional collaborations have created huge amounts of reliable information, and this information is free and easy to find.

"For example, there are giant, free product databases for every industry, sponsored by industry groups and created by volunteers, which list every single product, and every bit of information publicly available about them. No longer do we have to search through different company websites; everything is collected to make shopping and choosing items from neutral sources as easy as possible.

"Here�s another example. While much of the best professional literature for every industry is still available only to subscribers, you can also find it summarized elegantly and readably for free. You no longer have any excuse for not keeping up with what has been going on in your field, because the latest advances and best practices are summed up by collaborating volunteers, both at a high level of generality and in great detail.

"Also, the way we get our news has changed--again. It was in the late 90s and early 00s that the Internet began to eat into the profitability of news organizations. But in the last ten years it has been the widespread availability of collaborative news summaries which has completely disrupted the journalism industry. Some of us read only these news summaries.

"In education, at the primary, secondary, and university levels, the tools of the trade have undergone a complete revolution. Because there are tens of millions of high-quality, expert-approved encyclopedia articles, students and teachers can find reliable information on every topic they study amazingly quickly. These articles are integrated with original sources, all of which are available digitally now, so that every student at a school that pays its fees can have the resources of the largest library in the world from any computer terminal.

"And since there are comprehensive, regularly-updated reviews of the academic literature, these reviews have become the starting-point for the research of all graduate and postgraduate students, and indeed of professors themselves. Since academic fields have become so much more transparent and accessible, it is easy for academics to tell when a paper advances the field and when it does not; so standards are increasing.

"It is as if we are all now plugged into the World Brain, the global super-encyclopedia that H. G. Wells once dreamed of. We are all completely immersed in a world of the best available information. Most of us no longer spend much time finding or organizing information, because that is already done for us by an international army of collaborators; we spend our time, instead, directly accessing and personally making sense of the information. Finding and navigating it is now trivial and taken for granted.

"The revolution in collaboration that the world has undergone in the last fifteen years or so has also had profound social effects. Throughout history, the most influential, world-changing inventions brought the world ever closer together. Now, thanks to strong collaboration on the Internet, it is not at all unusual for us to be working together on a daily basis with people in India, Japan, and Egypt, and not just on nonprofit collaborative information projects, but in our conversations about our industries. Since the same incredibly useful information resources are available everywhere, and since we have created these resources together, these projects have brought us together, as professionals, into new global groups. Our awareness of each other�s intellectual and business traditions has become easy and expected, and in many ways we now have exactly what some, a generation ago, feared: a global monoculture. This is both bad and good, but it is what we�re stuck with; information shapes culture, and much of the world has access to much of the same information; so it�s not surprising if, in our professional lives anyway, we are moving increasingly along the same paths.

"The profoundly easy access to every sort of information now in 2017 has made us view information less as a proprietary thing, controlled by powerful elites, and more as a shared thing, like oxygen, open to whomever is interested in it. And since there is so little friction in discovering what is known in a field, innovation has accelerated even faster than it had been in previous generations.

"But arguably the most stunning impact of the collaborative revolution has been on India, China, and increasingly, many other countries in the developing world. In every corner of the world with access to the Internet, a new Enlightenment is taking place, because the intellectuals in each different country of the world are no longer confined to their own books and magazines and the few that they are able to import from abroad. They have access to the same information as the stock broker in New York City, and the college professor at Oxford, and the computer programmer in Berlin."

Well, that�s the end of my science fiction story. I would speculate about what the social impact of this "Enlightenment" might be on developing countries, but I am not a political scientist or a historian, and so I really can�t say anything about it without embarrassing myself. I do think, however, that access to information tends to increase standards of living, so it will probably be a very good thing.

Thank you very much.

Home