The Internet belongs to everyone

By Eric Lee

Published by Labour and Society International

Copyright 2000 by Eric Lee


This book is part of the series of Labour and Society International Discussion Papers. Some additional papers may be found online at LSI's website, The views expressed in this book are not necessarily those of LSI. We invite comment and discussion.


This book is dedicated to the memory of Michael Harrington


Thanks to David Post for allowing me to reprint his excellent essay, "Governing Cyberspace: Where is James Madison now that we need him?" Even though this appears at the end of this book, readers may profit from reading it first. Thanks also to Stirling Smith, Chris Croome, Ulrich Klotz and Peter Waterman for reading early drafts of this book and sending on their comments.

Who is this book for?

The Internet belongs to everyone and that is the theme of this short work. But this book is not for everyone.

This book is aimed at trade unionists and socialists, the members of what is called (at least here in Europe), the labour movement.

It is probably the first book written about the subjects like the domain name system for an audience that is a bit more familiar with collective bargaining and grievance procedures.

I'm aware that for many of the people reading this, the Internet is a great unknown. Most of you will probably use the net but few of you have a clue how it actually works.

I hope this books helps clear things up a bit. It addresses some of the issues related to Internet governance, though clearly not all of them. I'm aware that there are many important topics not covered here. I'm also aware that like most of you, I'm new to this subject. Therefore I welcome your comments and feedback so that I can improve future editions. (You can write to me at

To make things a bit easier, I've done up a list of acronyms used here, and that's one of the appendices. Also, the first time I use a new acronym (like DNS) I spell it out. (DNS is the domain name system.) And when introducing new and unfamiliar terms, I'll try to explain in plain English what they mean. (For example, the domain name system is the method by which bits move around the Internet. It is the means by which you reach a website and your email gets delivered to its correct addresses, using domain names like "" for example.)

I've prepared a list of website addresses and added this to the appendices -- if you want to know more about any of this, there's lots more online than I could fit into this book.

I've included a chronology at the end of this book and it should give you a broader sense of where we've been with the Internet and in particular, how much trade unionists have used it. (Did you know that more than a decade ago, long before Netscape, let alone Napster, long before most of us had even heard of the Internet, there were hundreds of trade unionists engaged in online conferences about things like new technology and the unions?)

Those appendices should help, but be aware that this is not a simple subject.

In writing these words, I am reminded of the way things once were. There was a time when the right of working people to vote was not yet widely accepted in most countries. Government was in the hands of experts: kings and landowners and the new capitalists; these were the people who supposedly had the skills, knowledge and ability to govern countries and empires.

It took a decades-long battle by trade unions and the new political parties that grew up around them to win the basic democratic right to have a say in who rules the countries in which we live.

Today we all spend more and more of our time online, as our lives grow more and more dependent on the Internet. It is becoming crucial that we once again raise the age-old demand of the labour movement: democracy, the right to be -- at least in part -- masters of our own fate.

That is what this little book is all about.

The online power elite

If the Internet is something important to us, we should certainly want to know who controls it, and particular, who determines its future.

By ownership of the Internet, we can mean many things. Obviously the cables on which the data travel are owned by corporations and governments, and the internetworked server computers are held by governments, corporations, universities, non-profit organizations, trade unions, and even individuals.

That all sounds very nice, making it seem like the Internet -- or at least all its physical components -- is owned by a very large and diverse group. But one very important part of this system, the root servers, is entirely in the hands of ICANN -- and possibly the US Department of Commerce. (There is a good discussion of root servers in the Appendix - "Governing Cyberspace: Where is James Madison when we need him?") That was true when the Internet was created thirty years ago and it is true today. I'll return to this point later.

There is another aspect to ownership of the Internet, and that is the part that concerns us here: Who sets the rules that govern this immense global network? Who decides its future direction? The simple answer is that giant corporations, usually US-based, are running the Internet today as they see fit, in accordance with their own interests (making a profit), with almost no input from civil society.

New world-wide institutions have been created to manage the Internet which, like the World Trade Organization (WTO), the World Bank, and the International Monetary Fund (IMF), are completely unrepresentative, undemocratic, and unaccountable. And yet they have the power to move the Internet in directions that may not always be the ones the rest of us would like to see.

Those institutions include the Internet Society (ISOC - founded in 1992), the World Wide Web Consortium (W3C - founded in 1994) and the newest of them all, the Internet Corporation for Assigned Names and Numbers (ICANN - founded in 1998). The three of them have a lot in common, not least of which the fact that all three have their headquarters in the United States.

They are all completely dominated by transnational corporate capital. It sets the policies under which the Internet and World Wide Web are governed. Their agenda differs not one iota from that of the corporate elite: for them, the Internet exists in order to help maximize profit.

Their deliberations are rarely reported in the mass media and only by accident do ordinary people discover their very existence. In their own defense, they will point out that they are "technical" bodies and therefore it is not necessary for loads of people to be involved. Really, thanks very much, they might say, but we don't need help from outsiders lacking in specialist technical skills.

But their own websites, in appealing for corporations to contribute money, make the opposite point. As ISOC put it bluntly on its site:

"What would it be worth to your organization to be able to shape the direction of the Internet and its related technologies? At the Internet Society, our public policy, legal, regulatory, and trade activities put you and your organization at the forefront of the Internet arena. As a member, you also hold a seat on the society's Advisory Council, which meets annually in conjunction with the Board of Trustees meeting and which has a significant voice in shaping the directions and initiatives of the society and thus the Internet itself."

"A significant voice in shaping . . . the Internet itself" -- now that's something I'd like to see not only for the corporations which have joined the Internet Society and similar bodies, but also for all of us. And by all of us, I don't just mean the techies, the geeks, the nerds, the wired, the clued-in, and so on. I mean everyone -- even those without Internet access. Especially for them. I think that one could even make the case that it is the people today who don't yet have access to the Internet -- in particular in the developing world -- who most need to have a say in how things develop. They have been excluded from the process since the Internet, in its original form as ARPAnet, was launched back in 1969 under the patronage of the US Department of Defense. They are still excluded today, more than thirty years later.

Their exclusion from the net since its early days has been expressed in many ways. Some of the things about the net that notoriously don't work are a function of its character as a network set up to promote the interests first of the US government, and now of transnational corporate capital.

For example, anyone who has tried to create (or sometimes, even read) websites in languages other than English may have wondered what kind of standards body set up the net in the first place.

I had the experience when living in Israel of being able to work in Hebrew on computers long before I could create web pages easily in that language. And there is still no single standard for reading Hebrew web pages. The reason for this is a simple one: the first version of Hyper Text Transfer Protocol (HTTP) -- which is the way web pages are retrieved by your computer and then displayed on your screen -- was written in such as way that only a handful of languages could use it properly. There was no problem at all creating a website in English, but if you wanted one in French or German and needed to include letters that are not used in English, you had a problem. This was eventually fixed, sort of, in later versions of HTTP. I'll have more to say about this later.

The source of this remarkable bit of information is not some band of wild-eyed radicals, but the Internet Society's own website. ISOC, like the other bodies governing cyberspace, is racing to fix the problem of internationalization of the web. Someone obviously figured out that there were potential consumers out there who wouldn't buy from all those nice new dotcoms because they couldn't read the websites.

So what kind of standards body creates a "world wide" web and forgets that not all languages use the same 26 English letters? The kind that isn't listening to the three fourths of humankind that doesn't speak English. If the institutions that launched HTTP had had the input of civil society, probably someone would have mentioned this little problem earlier on.

Excluded from the governance of cyberspace have been most institutions of civil society -- women's groups, farmer's organizations, religious bodies and, above all, trade unions. Finding a trade unionist involved in "shaping . . . the Internet itself" is as hard as it gets. They simply don't exist.

I don't want to look a the question of the governance of cyberspace in isolation.

The last few months have been characterized by an explosion of discontent at the secretive, undemocratic workings of the WTO, IMF and World Bank. Trade unions have played a vital role in expressing this discontent, and it was their involvement in the 1999 Seattle protests that made that event into a turning point.

This is not to say that all trade unionists have the same viewpoint on these matters. Some merely wanted a seat at the table, to have some input into the decision making process of the global bodies which regulate the world's economy. Others took a more radical view, seeing the WTO and similar bodies as being unreformable.

I think it's time to take a hard look at the secretive and undemocratic bodies which control today's (and more important, tomorrow's) Internet. This is important for all of civil society, but above all for the unions.

Why "above all" the unions? I think that trade unions are not only among the largest and most significant elements of civil society -- they are uniquely democratic and accountable to their memberships.

I admit that this is not true in every single case. There are unions that are not particularly accountable to their own memberships. There are unions that are controlled by the employers -- company unions. There are unions that are controlled by governments, such as the Chinese unions. There are unions that are controlled by gangsters, as was the case for some time in the Teamsters union in the USA. But those unions have been, by and large, excluded from the mainstream of the international trade union movement. Nearly all the unions which are today affiliated to the national trade union centers which in turn belong to the Brussels-based International Confederation of Free Trade Unions (ICFTU) are basically democratic, accountable to their members and serving their interests.

Because of this, the trade unions speak with unparalleled authority. They have the power to give a voice to the voiceless, and to make sure that the Internet will serve the interests of all humankind.

In this book, I intend to address a few issues related to the governance of cyberspace and in particular the role of the trade unions. I begin with the simple belief that the Internet belongs to everyone. This is an explosive idea. Do not underestimate it. Even the most radical critics of the way cyberspace is governed today rarely come close to saying what is obvious to anyone who looks at the way things are run today. It is time to say loud and clear that it is unacceptable that something as important as the Internet is regulated and controlled by bodies that are completely unaccountable to civil society.

Let's look at those bodies a bit more closely now.

The similarities between the bodies that regulate and control cyberspace and those which regulate and control the global economy are striking. They all tend to be US-based, controlled by people who come out of the corporate elite, and far removed from civil society.

Like the WTO, the World Bank and the IMF, cyberspace is run by an alphabet soup of bodies three of which I will briefly explore below. These include:


Let's begin with a closer look at ISOC, the oldest of the new global institutions that govern cyberspace. ISOC, which was founded in 1992, defines itself as "the international organization for global cooperation and co-ordination for the Internet and its internetworking technologies and applications". Its goals are lofty, and include the following:

development, maintenance, evolution, and dissemination of standards for the Internet and its internetworking technologies and applications;

growth and evolution of the Internet architecture;

maintenance and evolution of effective administrative processes necessary for operation of the global Internet and internets;

education and research related to the Internet and internetworking;

harmonization of actions and activities at international levels to facilitate the development and availability of the Internet;

collection and dissemination of information related to the Internet and internetworking, including histories and archives;

assisting technologically developing countries, areas, and peoples in implementing and evolving their Internet infrastructure and use;

liaison with other organizations, governments, and the general public for co-ordination, collaboration, and education in effecting the above purposes.

Ask ISOC how it actually works to achieve these goals, and this is the answer you get:

"The Internet Society operates through its international Board of Trustees, its Secretariat, its International Networking Conferences and Network Training Workshops, its regional and local chapters, its various standards and administrative bodies, its committees, and its volunteers. The Board of Trustees is headed by a Chair, and consists of 15 eminent individuals drawn from every region of the world - some of whom were instrumental in creating and evolving different components of the Internet and the technology. The Secretariat is managed by a President/CEO, with the assistance of several officers."

This tells us absolutely nothing. How does ISOC really work? Who, for instance, are the people at its very top? Let's look at the six members of the Executive Committee of ISOC's Board of Trustees. At first glance, it's not a bad list. There are representatives from different countries. There are some women. There is Vinton Cerf, the "father of the Internet", who, as we shall also see, sits on the ICANN Board as well. Presumably, these are the six most important people in the organization. But other than Cerf, I didn't recognize any of the names. Fortunately, in the spirit of openness that has characterized the Internet since its inception, all six have web pages so we can learn a bit more about them.

Vinton Cerf, the one ISOC leader whose name I recognized, is today senior vice president of Internet Architecture and Technology for MCI WorldCom.

Martin Burack is listed as ISOC's Secretary. Before coming to work full time for ISOC, Burack worked in various marketing management positions in strategic and data marketing for MCI over ten years.

Donald M. Heath is President/CEO of the Internet Society and according to his online biography is "a seasoned executive in the telecommunications, computer and software industries. He has held executive positions in development, operations, finance, sales, marketing and has worked in firms ranging from start-ups to very large multi-national corporations. He has served on the board of directors of a large-scale software firm for the past 14 years and acts as a business and financial advisor to a company engaged in the creation and development of sophisticated software development technology." He is today a Republican Party activist in Maryland. Earlier in his career, Heath "was Vice President, Data Marketing for MCI and was charged with creating MCI's data communications business".

Are we noticing a pattern here? Three Americans, all connected to MCI, one of the largest of the transnational corporations and one with a vital stake in how the Internet is managed make up half of ISOC's Executive Committee. Did someone say "conflict of interest"?

The other three are not Americans but all are business people too. Geoff Huston, from Australia, is Chief Scientist for Telstra Internet. Christine Maxwell is President and Publisher of Chiliad and has decades of business experience in the publishing field. Kees Neggers, from the Netherlands, has been Managing Director of SURFnet bv since its establishment in 1988. He is, apparently, the only one connected today to the academic and government bodies which were once at the Internet's core, and the mission of his company is "to develop and operate an advanced networking infrastructure for the research and higher education community in the Netherlands".

Anyone can join ISOC, according to the organization's website. But apparently no trade unions have done so. The annual membership fee for a non-profit organization can be as high as $25,000 or as low as $1,250. This has not deterred corporations from joining. ISOC is proud to include among its "gold" members (each one paying upwards of $50,000 a year for the privilege of membership in the club) the following:


Another key player in the governance of cyberspace is the World Wide Web Consortium. According to its website "the World Wide Web Consortium was created in October 1994 to lead the World Wide Web to its full potential by developing common protocols that promote its evolution and ensure its interoperability. W3C has more than 400 Member organizations from around the world and has earned international recognition for its contributions to the growth of the Web."

Who are those organizations? According to the W3C itself, "members include vendors of technology products and services, content providers, corporate users, research laboratories, standards bodies, and governments, all of whom work to reach consensus on a direction for the Web. These organizations are typically investing significant resources into the web, in developing software products, in developing information products, or most commonly in its use an enabling medium for their business or activity . . .. W3C Membership is available to all organizations."

Do trade unions meet any of these criteria?

Certainly unions are content providers. They often invest significant resources into their websites (or they should!) And for many unions, the web is increasingly "an enabling medium for their . . . activity."

So how many of the 437 organizations who currently make up the W3C are unions? Apparently, none. Not a single one. If you search their list for the word "union," you'll stumble upon the European Broadcasting Union. No trade union seems to have affiliated to the Consortium. I wonder why.

One reason might be the cost. The normal cost of joining is $50,000 a year. (Ouch.) But non-profit organizations are welcome to join at a much reduced rate -- only $5,000 a year, with three years payable in advance. Unions could become W3C members for only $15,000 each. No wonder they don't join. Even joining ISOC is a bargain by comparison. As a result, hundreds of corporations determine policy for the W3C and the unions have no say whatsoever.

A more likely reason than the prohibitive cost is that unions have probably never heard of the W3C.

The W3C seems to have a much looser structure than ISOC and its "team" consists of 64 people working from locations across the globe. W3C is hosted by the Massachusetts Institute of Technology Laboratory for Computer Science [MIT/LCS] in the United States, at the Institut National de Recherche en Informatique et en Automatique [INRIA] at various locations in France, and at the Keio University Shonan Fujisawa Campus in Japan.

Its management team is headed up by people with roots in the academic and research community; the chair is Jean-François Abramatic, who previously served as Director of Development and Industrial Relations at INRIA, which was set up in 1967 at Rocquencourt near Paris. It is a scientific and technological institute operating under the dual authority of the French government's Ministry of Research and the Ministry of Industry. The Director of W3C is Tim Berners-Lee, the inventor of the world wide web. He is also a Principal Research Scientist at the MIT Laboratory for Computer Science.

Though the W3C responds in its own way and in its own time to the needs of some parts of civil society (such as its recent campaigns to make the web accessible to the disabled and to internationalize a bit more), the mechanism for doing so is anything but democratic. It does, admittedly, sometimes work; the open email lists apparently are a relatively effective way for coders to get some input into new standards which the W3C drafts.

Nevertheless, there is certainly no body running the W3C that can be considered accountable in any way to civil society. In the end, the "team" does what its corporate sponsors, each one paying $50,000 a year for the privilege, want.


The Internet Corporation for Assigned Names and Numbers, founded in 1998, is the newest of the bodies set up to govern cyberspace. It has attracted a lot of attention lately both for its decision to allow the introduction of new top level domains (in addition to the existing com, org, net, edu, and so on) and also its ostensible "democratization" -- but more on that in a few pages.

For an excellent analysis of what ICANN is and why it is important, I've attached as an appendix to this book David Post's article, "Governing Cyberspace".

For now, I'd like to take a look at who actually runs ICANN, its Board. Keep in mind that the decision to create ICANN in the first place was taken by the US government in order to "free up" the control of the domain name system. But the body they created is neither democratic, nor representative, nor accountable. Like ISOC and the W3C, it consists entirely of academics and businessmen and it does on the net whatever serves the interests of capital.

Let's start with ICANN's Chairman, Esther Dyson. Dyson currently also chairs EDventure Holdings, described as a small but diversified company focused on emerging information technology worldwide. She sits on a long list of boards and is linked to, among others, Perot Systems. She began her career as a reporter for Forbes magazine (the one that jokingly calls itself "capitalist tool") and then became a securities analyst.

Several members of ICANN's Board also serve on other bodies we've already mentioned. For example, Jean-François Abramatic and Vinton Cerf have already been mentioned for the roles they play in ISOC and the W3C.

Other ICANN Board members come directly from the businesses that dominate all the institutions governing cyberspace today. Though we might not recognize the names of the individuals, some of the corporations they come from will surely be household words, such as Dun & Bradstreet, British Telecom, GTE, MCI, Bull, and so on. Among these individuals we find:

Géraldine Capdeboscq, Executive Vice President for Strategy, Technology and Partnerships at Bull, the French IT company.

George Conrades, Chairman and Chief Executive Officer of Akamai Technologies, Inc. and venture partner at Polaris Venture Partners, an early stage investment company. Conrades had been Executive Vice President and President, GTE. He previously worked as CEO of a little known company called BBN which played a crucial role in setting up the original Internet for the US military, as well as holding senior posts in IBM.

Phil Davidson, Head of BT [British Telecom] Group Standards in the BT Group Engineering and Technology Directorate. Frank Fitzsimmons, Senior Vice President, Global Marketing for Dun & Bradstreet, where he is responsible for the implementation of new global marketing initiatives in the areas of access systems, software and consulting partner marketing, Internet applications, electronic markets, and value-added products.

Greg Crew, Chairman of the Australian Communications Industry Forum Ltd., Chairman of the Australian Information Technology Engineering Centre Ltd., and a non-executive director of ERG Ltd. (Perth) and of Silicon Wireless Ltd. (California). He was Chief Executive Officer of Mercury Communications Ltd. (UK) (1993-95) and Chief Operating Officer of Hongkong Telecommunications Ltd. (1991-93). He was also Managing Director of Hongkong Telephone Co. Ltd. (1988-91). Previously, he occupied various positions in engineering and commercial management in Hong Kong, and in Australia with Telecom Australia.

Hans Kraaijenbrink, member of the Executive Board of ETNO, the European Telecommunications Network Operators association, located in Brussels. He is also Manager, European Policy and Regulation with Royal KPN N.V., the Netherlands, where he is responsible for European and international regulatory strategic affairs.

Another prominent group on the ICANN Board consists of a handful of lawyers and management consultants. These include:

Amadeu Abril i Abril teaches European Union Law, Competition Law, and IT Law at ESADE Law School, Ramon Llull University (a private University based in Barcelona). He also is an attorney-at-law specializing in distribution contracts, competition law and IT law. He has been acting as a consultant on Internet and e-commerce affairs to a number of European companies, most notably as Legal & Policy Advisor to Nominalia Internet SL, a domain-name registrar.

Jonathan Cohen is the Senior, Managing Partner of the Shapiro Cohen Group of Intellectual Property Practices, located in Ottawa, Canada. He has practiced in all areas of intellectual property (IP) law since 1971, and has lectured and written extensively both in Canada and internationally on various aspects of trade-mark law, including, more recently, domain name issues Ken Fockler is the President of Tenac Consulting, a company he founded.

Eugenio Triana is an International Management Consultant on telecom policy, space and satellite systems, copyright and intellectual property rights in Madrid.

Pindar Wong is the Chairman of the Asia & Pacific Internet Association, Executive Committee Chairman of the Asia Pacific Regional Internet Conference on Operational Technologies, advisor to the Asia Pacific Networking Group and member of the Editorial Advisory Board of Cisco Systems' Internet Protocol Journal. He is also the Chairman of VeriFi (Hong Kong) Ltd., which is described on the ICANN website as "a discrete Internet infrastructure consultancy".

Finally, five members of ICANN's governing body come from academia and the research community which has been closely associated with the Internet since the days that it was founded as a secret project of the US Department of Defense. Among these are:

Michael M. Roberts, who serves as President and CEO of ICANN, is "a policy consultant in the field of Internet technology, services and product development, with a specialization in research and education . . .. He recently retired as Vice President at EDUCOM, a consortium of 600 universities and colleges with interests in information technology, where he was responsible for networking and telecommunications programs, including the development of public policy positions in information technology on behalf of EDUCOM members." He was the first Executive Director of ISOC before coming over to ICANN.

Jun Murai, Professor at the Faculty of Environmental Information, Keio University (Japan), who also serves also on the Board of ISOC.

Robert Blokzijl, who currently works for the National Institute of Nuclear Physics and High Energy Physics (NIKHEF).

Alejandro Pisanty, Director of Computing Academic Services at UNAM, the National Autonomous University of Mexico, in Mexico City, Mexico.

Linda S. Wilson is president emerita of Radcliffe College, presently on sabbatical leave after serving as president for a decade. Previously, she was vice president for research at the University of Michigan, and served in the senior administrations of the University of Illinois and Washington University, St. Louis.

As we go over the names and the illustrious careers of the members of the bodies that today govern cyberspace -- ISOC, W3C and ICANN -- we find some overlap, but more important is what we don't find. We don't find representatives of civil society, and in particular we don't find trade unionists. Trade unions are the largest organized force in civil society, but they have no say whatsoever in these bodies.

It is time for this to change -- a fact that is recognized by nearly everyone. As a result of this recognition, ICANN announced that it was going to become democratic; members of its Board were going to be elected by Internet users themselves.

Actually, not all members of its board. And not by all Internet users.

Still, this sounds better than the current situation. Critics, however, were swift to point out the problems with ICANN's own "democratization".

But before going into their criticisms, let's just note that even those proposing greater democracy are themselves unrepresentative; where are the trade unions and other democratically accountable bodies of civil society in this process?

Take the so-called Internet Democracy Project, whose statement on the eve of ICANN's Yokohama meeting in July 2000 (which appears as an Appendix to this book) is basically a call for greater democracy and openness. It was entitled "Civil Society and ICANN Elections".

Of the original fifteen signatories to the document, who come from all over the world, not a single one is identified as a trade unionist. They come from such organizations as the American Civil Liberties Union, the Association for Progressive Communications, universities in South Africa and the US, and various groups of socially minded professionals, such as Computer Professionals for Social Responsibility. Not a single trade union nor identifiable trade unionist is to be found. Were unions involved in drafting such a document? There's no sign of that. Were unions asked to support the initiative? Presumably yes -- once it had been drafted.

This is not intended as criticism of the Internet Democracy Project (IDP). On the whole, it all sounds like a very good idea. But it must involve democratically accountable representatives from trade unions if it is to work.

Let's take a look at a couple of their proposals to see how unions would fit in -- what we would support and what not, and perhaps try to imagine what this document might have looked like had unions been involved in its authoring.

For example, one of the proposals reads as follows: "The DNSO [Domain Name Supporting Organization] should recognize new constituencies, including an Individual Domain Name Holders constituency, a developing countries constituency, and a small business constituency."

It's very nice to single out, for example, these three examples of groups that would seem to have special needs. Not coincidentally, among the 15 original signatories are representatives of something called the "Individual Domain Name Holders Constituency (USA)" and developing countries. But what about the unions?

Let's think amount this for a moment. What are the special needs of small businesses? Presumably the fact that they have fewer resources (less money) than big businesses. If there is a need for the DNSO, which is an ICANN subdivision, to recognize new constituencies that have special needs, whatever that means, I think that we might have wanted to include unions on the list. (Remembering that there are a lot more workers than small businessmen in the world today.)

The people at the Internet Democracy Project are well aware of the completely undemocratic way in which ICANN proposes to elect the five "at large" members of its Board. It allows anyone to register, but by the 31 July 2000, when the registration period ended, not even one tenth of one percent of the people who use the net have done so. Put another way, more than 99.9% of Internet users aren't registered to vote in ICANN's elections. (ICANN called this -- in all seriousness -- a great success. Furthermore, they declared that "the overwhelming number of registrations produced significant logistical and financial problems for a system that was established and intended to deal with fewer than 10,000 registrations.")

But the IDP's solution is hardly a radical one.

They write: "Internet users in many developing countries have Email but not web access. ICANN membership should be possible (and easy) with just an Email connection." At first glance this sounds very nice, very inclusive, and of course unions would not oppose anything that allows more people, particularly people in developing countries, to have a say in the process.

The problem is that this does not go nearly far enough. The Internet is for everybody; everybody should have a say in how it is run. This is, I think, the ABCs of how progressive people, democrats, trade unionists should look at the issue of how cyberspace is governed.

Instead of merely demanding the right for a few people in developing countries who might have email access but not web access -- instead of fighting merely for them, why not demand a much more inclusive approach? I think that the billions of people who are not yet online but who will eventually make up the vast majority of Internet users need to be included in the decision making process even now.

The only reason why some interests of the Chinese people might be represented in some way today -- for example, the internationalization of web authoring standards to allow non-Latin characters to appear on websites -- are a result not of the pressure from a thousand million Chinese people, but the pressure from global businesses eager to sell to the Chinese market.

A democratically governed Internet would give those people a far greater say than they have today.

Is it practical to hold global elections not based on email and the web, open to everyone? Probably not. There are certainly no such elections held for the United Nations, for example, though it would be good if there were. But trade unionists would probably argue for the principle of an inclusive decision making process in which many more voices could be heard. And the best way to do that now is to replace the undemocratic decision making process in which those people who happen to have access to the web or email and who stumble upon groups like ICANN with a process by which the disenfranchised millions get to have a say.

One way to do that is to involve representative, accountable, and democratic bodies in the governance of cyberspace.

Those bodies would include (but not be limited to):

This is a far more radical and more democratic approach than merely begging ICANN to allow a few thousand people in developing countries to vote by email.

And it expresses the principle which must guide a democratic approach to cyberspace which is that the Internet belongs to everyone.

How would we go about securing the participation of these groups in the governance of ICANN -- and ISOC and the W3C for that matter?

Fortunately we have one working model in the form of an international body which does recognize the need to ensure participation in the decision making process by ordinary working people and their unions. That model is the International Labour Organization (ILO) and it is probably the only part of the UN system that ensures a voice for workers.

The ILO is run on the basis of tripartitism, with government, business and unions all having a voice. It is obviously not an ideal system. But it works.

Opponents of any proposal to make bodies like ICANN less like the WTO and more like the ILO will hear that the ILO deals with labour matters, so it's reasonable to include unions. But what stake do unions have in the governance of cyberspace? I'll try to give that question an answer in the next chapter, but for now lets say this:

The roles of ICANN and the ILO are different. My proposal is simply this -- if the ILO can guarantee one third of its control to the workers, what can ICANN offer to the trade union movement? A single seat on its Board? Two seats? Or no say at all in the way cyberspace is governed, which is where things stand today.

Once we accept in principle that unions need to be represented, as do other components of civil society, the question of how much representation is a matter for bargaining.

In any event, an ICANN, ISOC or W3C Board on which there sit representatives of trade unions will be infinitely more democratic and accountable than it is today.

Labour's stake in the net

I have been repeating a demand here for unions to be given a much more active role in the governance of cyberspace. Here I want to back that up with a discussion of three issues:

Unions have been using the new technology is an innovative way for some twenty years now.

Trade union members today make up a huge segment of the Internet population, with well over 45 million of them online.

At least one of the issues being discussed by ICANN today vitally concerns the unions -- the possible introduction of a new "union" top level domain.

Twenty years of trade union use of the net

I have been writing about unions and the Internet since 1993. That was even before Mosaic, before Netscape, before anyone ever uttered the words "e-commerce". And I'm not even one of the pioneers of labour's use of the net -- by the time I began writing about it, extensive work had already been done around the world. (For a detailed account, see the Chronology at the end of this book.)

The first published suggestion that trade unionists might be able to use networked computers for such tasks as online distance learning was made back in 1972 by Charles "Chip" Levinson, a visionary trade unionist, in his book International Trade Unionism.

I had the chance to meet Chip Levinson several times in the late 1970s. Unlike most trade unionists back then, Levinson was single-mindedly focussed on multinational corporations. He had already discovered a process we now call "globalization" and he saw this as both a threat to the unions and a great opportunity. It was a threat because multinational companies could easily shift production around the world in search of cheaper rates of pay and environments hostile to unions. Strikes could be broken more easily than ever before. Unions negotiating at national level with global companies had little leverage.

But Levinson also saw an enormous potential for the creation of what he called a "countervailing power" to those same multinational corporations -- global trade unions. And few were as well placed as Levinson to note these developments. He was general secretary of a Geneva-based international organization of chemical workers, uniting all the major unions of chemical workers around the globe.

Levinson was so far ahead of his time that it is staggering. And not only politically. He wrote about the possibility of unions using computers to conduct worldwide research on multinational corporations. He was excited about the possibility of using computer networks to conduct trade union courses. He wrote about things which we now understand are possible, but which have still not yet been done.

A few years after his book appeared, in 1981, a province-wide network of teacher trade unionists was functioning in British Columbia, Canada using primitive laptop computers equipped with modems. That network helped the union survive a period of unrelenting attack by a hostile provincial government.

By the mid-1980s, email was being used by a couple of the international trade secretariats -- global organizations of national trade unions in specific sectors, including Levinson's own chemical workers international. The Canadian Union of Public Employees, the largest union in that country, broke new ground with its Solidarity Network (Solinet), which was the first data packet switching national network in Canada.

Unions were also among the first to launch sites on the World Wide Web. In 1994 the Tel Aviv-based International Federation of Workers Education Associations was the first international labour body with a site of its own. (I know about that one because I designed the site myself.)

In other words, trade union use of computer communications is nothing new. It has a history going back almost to the beginning of the Internet itself. Yet even as the governance of cyberspace became an issue, unions were not invited to participate in that process, nor, it must be said, did they demand the right to participate.

While many unions have created websites and many individual trade unionists have taken an interest in some of the new issues raised, on the whole the conservative culture of trade unions has stood in the way of the active involvement of these organizations in the decision making process. Trade unionists have tended to avoid complex issues connected to the new technology with only a few exceptions.

One of these has been a world-wide campaign launched a couple of years ago by Union Network International (UNI), the largest of the international trade secretariats, representing 15,500,000 trade union members in over 900 individual unions around the globe. UNI's campaign is called "Online Rights for Online Workers". The campaign has focussed both on the right of access by trade unions to the electronic workplace and also the rights of individual workers to unmonitored, free access to the net. But even this campaign has more to do with conventional workplace issues than the new issues raised by a networked economy.

Trade unions have by and large been outside of the debates that have rocked the Internet in recent years -- such as freedom of speech on the web, the protection of intellectual property rights, the open source movement. Some of these terms -- such as "open source" -- will be unintelligible to most trade unionists.

This is unfortunate, as the tens of millions of trade unionists who now use the net have a vital stake in some of the issues now being confronted by those very same institutions on which they have no representation.

There are a lot of trade unionists online

Members of trade unions constitute one of the largest groups in cyberspace. How large is that group? No one knows, of course. But let's try to make an educated guess.

By the summer of 2000, estimates about the size of the total population of Internet users reached nearly 333,000,000. These are of course only estimates. There could be only a hundred million online, or even half a billion. And of course, a lot depends on what you mean by "online".

A lot of people probably have had access to the net once or more without owning their own personal computers. A lot have access to email but not to the web. There are many who use the Internet for very narrow purposes at work, but have never in their lives visited a website because they wanted to learn something or do something. There are plenty of people who have been given access to the net by their workplaces, with their own email addresses even, but have never used them.

The number of trade unionists online is a certain percentage of the hundreds of millions of people who are now online. While no one has done a census of Internet users, there are a few things that would seem to make sense and can serve as a guide.

In the advanced industrial countries, trade unionists will tend to have greater job security and higher salaries than non-unionized workers in the same sectors of the economy. In other words, they will be more likely to have the disposable income necessary to purchase computers and other costs associated with going online. In the US, at least, studies show that trade unionists do have a greater likelihood of owning a personal computer than non-trade unionists.

Therefore, it is not unreasonable to assume that the number of trade unionists online is at least as high as the percentage of trade unionists in a given society. That means that in a country where 10% of the workforce is unionized, the number of trade unionists online will be at least 10% of the total number of Internet users in that country.

Using that way of guessing, and recent figures from NUA Internet Surveys, an Irish company which surveys the Internet every month, and the ILO, we find that in a dozen industrialized countries, including some with very high and others with very low trade union density, there would seem to be something on the order of 45 million trade unionists online (see table below). That's 45 million online just in those 12 countries; extrapolating to the Internet as a whole, the figure might reach 60 million.

Obviously these are not exact figures -- not only are the NUA and ILO numbers not much better than educated guesses, but the idea that one can simply apply the percentage of trade union members to the number of people online is not proven. But here are what the numbers tell us today:

CountryTrade union densityNo. of Internet users Trade unionists online
Denmark68.2% 1,900,000 1,295,800
Finland 59.7% 2,150,000 1,283,550
Norway 51.7% 2,200,000 1,137,400
Canada 31.0% 13,280,000 4,116,800
Germany 29.6% 15,900,000 4,706,400
Australia 28.6% 7,000,000 2,002,000
UK 26.2% 19,360,000 5,072,320
New Zealand 23.2% 1,270,000 294,640
Japan 18.6% 27,060,000 5,033,160
USA 12.7% 134,200,000 17,043,400
France 6.1% 9,000,000 549,000
With the notable exception of the United States, there is a correlation between trade union density and Internet density in countries. Countries which have achieved the affluence of, say, Sweden, tend to simultaneously have very large numbers of people online -- and very large and powerful trade unions.

So in liberal capitalist democracies like Denmark, Canada and New Zealand, you have loads of trade unionists and loads of Internet users. In backward, poor, dictatorships like Iraq, Burma and North Korea, you have almost no Internet users and almost no trade unionists.

This natural trend which has brought trade unionists online in their millions has been helped recently by unions which have aggressively pushed to get their members online. In Sweden, Australia and the USA, there are well-financed, well-publicized initiatives organized by the unions to give their members PCs and Internet access at reduced cost. The Swedish case is especially interesting -- one of their national trade union centers is reputed to be the largest single purchaser of computer equipment in the country. The Australian unions were divided over how to boost the number of trade unionists with net connections, and eventually both the Australian Council of Trade Unions and the New South Wales Labour Council launched initiatives which now compete with one another. There is no disagreement about the need for trade unionists to have computers and Internet access, only about the details of the way in which this is to be done.

Even the recent well-publicized announcement by Ford that it would be providing computers and Internet access to all its employees around the globe was initially a trade union initiative. It was the vision of the United Auto Workers (UAW) in the US, and not Ford management, that led to this striking announcement.

So how many of the 330 million online today are trade union members? Lots -- probably more than the 45 million in the table above. Possibly 60 million. Maybe even more. And the number is growing, as more and more working people come online. This makes sense: business people, professionals, and the like have had Internet access for some time now and there is no longer growth of Internet use among the rich.

How many of those tens of millions of trade unionists are involved in the governance of cyberspace, helping to set the rules under which the Internet works? None.

A cyber union label?

On 14 July 2000, while ICANN's Board was meeting in Yokohama to discuss, among other things, the introduction of new top-level domains (TLDs), they received an "expression of interest" which was duly numbered number 26. It came from an unusual source: the Brussels-based International Confederation of Free Trade Unions on behalf of a consortium of international trade union organizations.

The paper began with a disclaimer, stating that "the policies mentioned in this paper are likely to be features of an eventual proposal for a trade union TLD" but this proposal "is meant only as a guideline" and that possibly "some or all policies mentioned could be altered considerably, following a full discussion among the sponsoring group on the content of a proposal."

The document began by explaining who was submitting it to ICANN. This is not a simple chore when describing organizations that are not only unfamiliar to the businessmen and academics in ICANN, but are not even well know to trade unionists. How many ordinary working people have ever heard of the ICFTU or any of the international trade secretariats (ITSs) which joined it in this proposal? Probably very few.

The ICFTU, it begins, "is submitting this expression of interest, on behalf of a group of international trade union organisations and their national affiliated organisations." By adding the second phrase -- "and their national affiliated organisations" -- the ICFTU was claiming to speak on behalf of itself, the national trade union centers and individual unions in each country. It was an unusually long list, and the so-called sponsoring group included the following eleven international trade secretariats:

In addition to these, the sponsoring group includes the Trade Union Advisory Committee to the OECD.

To make clear to ICANN just how unusual was this proposal, the ICFTU added the following description of itself:

"ICFTU has 216 affiliated organisations in 145 countries and territories, and represents 123 million workers, 39 million of whom are women. The ICFTU membership includes, for example, the AFL-CIO (USA), CLC (Canada), DGB (Germany), TUC (UK), FNV (Holland), COSATU (South Africa), LO and TCO (Sweden), CUT (Brazil), KCTU and FKTU (South Korea), ACTU (Australia). In addition to the sponsoring group listed above, all of these national trade union centres are closely associated with this expression of interest."

This was an extraordinary development. Democratically accountable organizations representing literally tens of millions of people had joined together to ask ICANN to do something. This had probably never happened before. After all, not many people were behind the proposals for the other suggested top level domain names, such as museum, shop, biz, and -- this is my favorite -- sucks.

One would think that ICANN would have no choice but to listen to such a proposal, though it need do nothing until a formal proposal is submitted in the fall.

The proposal then turned to the structure and purpose of a proposed union top level domain. This, for us, is the really interesting bit.

It begins by answering one of the obvious questions which is, would the union domain name be open to everyone who wants one (like com or org) or would be a chartered domain, one in which a particular body would decide who gets to use the domain name. The ICFTU has clearly opted for the latter solution -- a chartered domain.

As for the domain name itself, the preferred string "would almost certainly be union", says the proposal, though interest is shown in the possibility of a multilingual domain, one in which "alias" TLDs are recognized, such as syndicat or sindicato.

This is an interesting point. When the first top level domains were launched some twenty years ago, they were all given English language names -- indeed, American English names. Com, org, edu, mil, gov -- these are all shortenings of English words. And mil means only US military, gov means only US government, edu is only for US universities, and so on.

The rest of the world got country code Top Level Domains (known in netspeak as ccTLDs -- don't you just love these acronyms?), and some of these got a taste of the Anglocentric character of the net. For example, the domain name given to the USSR, before that collapse, was "su" for Soviet Union. Now of course no one in the Soviet Union would have thought of using "su" for its country code -- the country's own acronym was, as everyone knows, "CCCP", or in English transliteration, "SSSR". Other country codes more accurately reflected local flavor, such as de for Germany (rather than gr). One surviving oddity is the ccTLD for South Africa -- za, obviously a relic of the days when the country's dominant language (and culture, and everything else) was Afrikaans.

So the ICFTU's suggestion that its domain might have to be multilingual -- and this at the very beginning of its first formal proposal -- is extraordinary. It shows the enormous gap between the thinking of an international trade union organization and the thinking in ICANN and its predecessor body. It is also very different from the early proposals made in the US -- not by trade unionists -- for a union domain name. They did not take into account the possibility that some trade unionists might not be comfortable with an English-only TLD. But the ICFTU did.

Immediately after this bombshell came a second. The ICFTU was asking ICANN to recognize already at this stage what is, and what is not, a union.

The proposed union TLD "would be open to registrations from representatives of trade unions which are independent of outside control and have a democratic structure".

Note the two qualities required by this sentence -- independence and democracy. This is what stands behind the word "free" in the name of the International Confederation of Free Trade Unions: the recognition that there are organizations which claim to be trade unions but are not. The ICFTU was saying right off the bat that this domain name will be limited only to those unions which we believe are real unions, and not to all the fakes. In other words, the domain name would become a kind of cyber union label, certifying that any site which bears this label is a real trade union.

As no one has the kind of control over the world's languages as ICANN will have over domain names, until now any group, anywhere could call itself a trade union. With the kind of definition the ICFTU is proposing, for the first time in the 150 year old history of the modern trade union movement, we will have a single, global, recognized definition of what constitutes a bona fide, independent, democratic trade union. Every one that meets that definition will be allowed to proudly display that fact on their website.

More on this in a moment, because at this point in the proposal the ICFTU decides to explain the five reasons why the world's trade unions would want a union TLD. And they are:

"To provide a strong and clear identity for workers' organisations on the Internet."

That sounds reasonable enough; after all, most unions share the "org" TLD with zillions of other groups few of which have any connection to unions. This would sort out the unions from all the others.

"To facilitate the efforts of employees to find and contact trade unions in their country, sector, or enterprise."

This means, of course, that the union TLD would have to be used in conjunction with the names of the country, sector or enterprise. This assumes a one-to-one correlation in each case, which doesn't exist. This is a bit of a problem - and the ICFTU is aware of this.

For example, the ICFTU has two affiliates in Korea, two national trade union centres -- the Federation of Korean Trade Unions (FKTU) and the Korean Confederation of Trade Unions (KCTU). They mentioned these in their proposal. The differences between them, despite similar names, are substantive. They do sometimes cooperate. If one were using the union TLD to point to trade unions in Korea, which one would it point to?

Even more common is the presence of many different unions in a single enterprise. In the United Kingdom, there are a number of different and competing teachers unions. Who would decide which of them would get the coveted uk.teachers.union or domain?

Yes, the obvious solution would be to get unions to cooperate and have the union domain point to all the relevant sites. But this does not seem at all simple.

"To help internet users identify bona fide trade union organisations, as distinct from bogus unions such as government-sponsored labour fronts, and company-controlled unions."

Now we come to meat of the matter. This is really what it is all about; a global trademark, enforced by ICANN: a cyber union label. It really is a problem that so many groups can pin the label "union" on themselves and get away with it; this is one solution. (Though one wonders how many of those groups -- those phoney unions -- have a presence in cyberspace.)

For example, the ICFTU's chief rival -- over a more than 50 year period -- for the claim of representing the workers of the world is the World Federation of Trade Unions (WFTU), which represents mostly the same "government-sponsored labour fronts" that the ICFTU warns about. WFTU also has some bona fide trade union affiliates as well. And just to complicate matters, some ICFTU affiliates have good working relationships with "government sponsored labour fronts", such as the official state-controlled unions in Cuba (which are not affiliated to the ICFTU).

What happens when the state-controlled Cuban unions apply for a union TLD? This too is not at all a simple issue, not even within the ICFTU, and I'll return to it.

"To form part of the ongoing international effort to bridge the 'digital divide', by building meaning and utility into the Internet for workers, regardless of country, or economic status."

We all want to bridge the digital divide, meaning that it would be great to provide the hardware, connections and above all training to allow the working class and poor around the world to use the net. Presumably the ICFTU means that website addresses become more "meaningful" if they have union at the end of them rather than org.

"To facilitate employee and public access to a wide variety of union-sponsored services, including apprenticeship and training programmes, health and pension benefits, family and community services, etc. "

Again, this seems to be saying that the creation of a "union" TLD will make use of the Internet easier for workers.

How exactly would a union TLD facilitate access? Presumably by allowing people to guess domain names.

In other words, let's say I work for General Motors in Michigan. I want to find my union's website. Maybe I'm not even a union member. I don't know the name of the union, so I couldn't possibly guess "", which is the real and current address. So instead, I type, "gm.union". Or is it, "generalmotors.union". Or maybe "general-motors.union". Of course with the ICFTU running things in the proper internationalist way, it might well be "generalmotors.usa.union" or "usa.generalmotors.union". Though of course if one gets stuck, you could always try "generalmotors.sindicato". That might work.

I think in the time it would take our autoworker to guess the correct name, he might have found the site more easily using a search engine like Google or AltaVista. (This is, after all, how search engines make their money.)

All this presumes that workers have learned that there is such a thing as a union TLD and that they can try to guess addresses this way. Of course the UAW could help by publicizing widely the fact that there is now a "generalmotors.union" website, but how is that any easier to do than publicizing ""? The more I think about this argument -- that a union TLD is someone going to make it easier for unorganized workers to find their union on the web -- the less reasonable the argument sounds.

I'll go further than that. I'm not convinced, despite all the domain name hysteria that has resulted in some "valuable" domain names being sold for millions, is anything more than that -- hysteria. Let's look at some of the most popular sites on the web -- what are their domain names? Yahoo, Lycos, MSN, Amazon, E-bay, ICQ, Geocities. Ten years ago, could any of you guessed what any of those words meant?

When I first began buying books on the net I did it through a website with the "valuable" domain name of I didn't find it because I typed "books" and hoped for the best -- I did it because it appeared as a link somewhere. Or I read about it in a magazine. As I recall it was a bookshop which charged practically nothing for shipping books across the Atlantic from the USA. It had the delightful name of "Book Stacks Unlimited" and its owner was Charles Stacks, if you can believe that. Did this domain name, so easily guessable, help them when coming up against Not one iota.

So I think that the importance of domain names has been hyped a bit, and as the examples I've show may indicate, you can pick a domain name out of a hat and if you properly promote it, you'll have the most popular, easy-to-find website on the planet.

We return now to the union TLD and how the ICFTU proposes that it be managed. The registrant database -- meaning the lucky websites that get to wear the new cyber union label -- "would most likely be managed by the sponsoring group" -- meaning the ICFTU, the 11 international trade secretariats, and the Trade Union Advisory Committee.

(The World Federation of Trade Union can forget about its chances of having a say here, which is no doubt a good thing.)

As the ICFTU and ITSs have little experience (meaning, no experience at all) in running a domain name registration system, they do admit that "the arrangement would probably involve the creation of a review body comprising the group's 13 organisations. The body would be empowered to accept or reject an application to become a registrant. A review procedure would be set up, to appeal decisions when and where necessary."

The ICFTU says that organizations which will want the coveted union TLD will have to meet two criteria:

As the proposal notes, "these are the necessary conditions for the existence of representative workers' organisations capable of engaging in collective bargaining, within the definitions and jurisprudence established by the United Nations' International Labour Organisation."

But these are not the only criteria. Other ones, currently unspecified, might also be applied. The proposal goes on to say that the ICFTU intends "to explore further refinements to ensure practicality and application. Additional criteria may also be identified and considered as well." One wonders what those additional criteria might be. For example, would ethnically exclusive unions be banned from using the union label?

Lest anyone think that the proposal aims to be exclusive, the ICFTU then goes on to say that "non-profit organisations, such as apprenticeship and training organisations, affiliated to approved unions, would also qualify as registrants." This seems like quite a hornet's nest -- what exactly constitutes a non-profit organization which is affiliated to a union? For example, are workers education associations, often funded by the unions, to qualify? How about producer and consumer cooperatives? And if these meet the criteria, why not those political parties which are controlled by unions? One would think that training organizations wouldn't need a union domain name -- wouldn't it be enough for unions to link to them?

The ICFTU goes further in its inclusivity, explicitly saying that you don't have to be affiliated to it or one of the other sponsoring organizations to get the union label in cyberspace. "Through the worldwide networks of the ICFTU, and its review body, a system would be devised to screen potential registrants. Applicants not affiliated directly or indirectly to one of the organisations in the sponsoring group would not be excluded from consideration if their organisation is known to be free and democratic." And, presumably, meets the other criteria not yet specified, whatever that is.

Now we come to the nightmare scenario. What happens when the Chinese, Cuban, Syrian, Iraqi and North Korean unions apply for their union domain names, the review body says no, and then the ICFTU-controlled appeals court says no as well?

The result will certainly be that WFTU and its various affiliates will approach ICANN and demand their right to run their own domain name for "real" unions. (Surprising they haven't done this so far, but then again, WFTU doesn't even have a website.) And when ICANN gets the letter from Prague, or wherever WFTU is headquartered these days, saying that it represents enormous numbers of trade unionists around the globe, who will get to decide which group gets to run the unions domain?

ICANN decides. Or more precisely, ICANN's Board, filled with its top management corporate types and academics, with not a single trade unionist, free or otherwise, anywhere in site.

The ultimate arbiter of who constitutes a real union will not be the ICFTU but the body it currently appeals to -- ICANN. Unlike in the previously mentioned International Labour Organization, where trade unions have a say, here the unions are submitting themselves to the authority of an undemocratic, unaccountable body in which they have no influence. And that body will be the one to decide, in the end, who gets the wear the cyber union label.

This is one reason why the democratization of ICANN should be a central concern for unions.

Now we come to the technical bit -- how exactly does a fifty year old trade union whose experience with the Internet consists basically of maintaining a website intend to actually run a domain name registration system? The answer is, it doesn't. It plans to subcontract that part out. "A contractor would most likely be responsible for 'registry operation'," says the ICFTU statement. "Its responsibility would include the technical management and registration of second-level domain names as well as the billing of approved registrants for those registrations."

The division of labour envisioned by the unions making this proposal is an interesting one.

The "technical management" of the unions domain will be in the hands of an unnamed sub-contractor. They will handle billing, for example. (So much for the idea that a unions domain might be, like the "us" country code top level domain, free of charge.)

The role of the sponsoring organizations will be entirely political -- they will decide who gets to wear the union label. Presumably they would also get a share of the registration fees as well. They'll certainly need that to cover the costs ICANN is going to charge them for the privilege of running a top level domain. Merely to look at such a proposal for a TLD -- which might be rejected -- ICANN will be charging $50,000. As ICANN says on its website, "the application fee is non-refundable and ICANN's only obligation upon accepting the application and fee is to consider the application."

As for those charges for getting a union domain name for your website, the ICFTU proposal doesn't give anything away. However there is one concrete suggestion -- "We would also expect that, at least in the early days, trade union organisations from the wealthier industrialised countries would make up a large proportion of the registrations, since the majority of workers with access to the Internet is currently based in these parts of the world. We foresee a strategy to bridge the 'digital divide' and promote greater use of the web by unions from developing countries, by making the cost of domain registrations considerably lower for registrants from poorer nations."

Is it reasonable to assume that trade unions in the wealthier countries will be the first to adopt the new union TLD? Maybe not. It might be even more reasonable to assume that many of these will stick with their existing names and that actually most purchasers of the union TLD will be new to the web -- keeping in mind that the vast majority of unions have no website today.

The idea that the labour movement's contribution to overcoming the digital divide will consist of discounted domain name prices for unions in developing countries is ludicrous. No one can seriously believe that the cost of a domain name is a real obstacle to a union going online. With domain name charges plummeting around the world -- here in the UK, the domain name is usually offered up free by web hosting companies -- this proposal is not even a drop in the bucket.

This brings me to the whole question of costs. ICANN has already announced that it is going to charge $50,000 for the privilege of having a chartered top level domain which needs to be paid by whoever wants to manage the domain. Presumably if you get to manage a domain name like "shop", you can then charge whatever you want to anyone who wants to have ".shop" at the end of their website or email address. Where the consortium of unions that is sponsoring this proposal plans to come up with the tens of thousands of dollars needed for this has not yet been specified. And even if the unions had this kind of money lying around, one wonders if it could not be put to better use.

There are also potentially enormous hidden costs to having a union top level domain. Thousands of trade union websites would be encouraged to begin using the new domain (otherwise, why have it?). Indeed, this is what the ICFTU is saying -- and it presumes that many of those unions who currently have websites will agree to pay for the union TLD as a way of subsidizing those in the poorer countries who will not be able to afford this.

They would have to publicize the new domain to their members and the media, and change everything printed that currently has the old domain. It's a bit like changing telephone numbers -- and this can be an enormous cost for businesses.

Remember that we're not only talking about websites. When unions change their domain names, they'll have to change their email addresses too. No one knows how many trade unionists have email addresses with the names of their unions today, though certainly thousands of trade union staffers already do. These will have to be changed as well.

Of course none of this need be done immediately -- obviously for a long time to come, unions would keep both their old and new domain names. But that's not the point -- a massive cost, both in money and effort, will be involved in switching over thousands of unions to a new domain name.

But let's leave that for a moment and return to the ICFTU proposal.

At this point in the ICFTU proposal there is finally a reference to the practicalities of resolving probable disputes. "In cases where it is alleged that the names or acronyms of existing trade union organisations have been registered in bad faith, we expect that we will find the machinery of the Uniform Domain-Name Dispute-Resolution Policy (UDRP) adequate to handle disputes," it says.

One wonders how many trade unionists are familiar with the UDRP. My guess is that for most of you reading this, this is the first time you've ever heard of it.

In checking out the UDRP pages on the World Intellectual Property Organization (WIPO) website, I came across a ruling made very recently -- 20 July 2000 -- concerning someone who'd registered national variations of the domain name "". The person who registered the domains was not a campaigning organization and was apparently trying to sell the domain names to Wal-Mart. The UDRP panel -- which consisted of one person -- ordered that the rights to these domain names revert to Wal-Mart. The original website was allowed to stand. As one who believes that Wal-Mart, one of the most anti-union companies in the world, in fact sucks, this decision is not good news.

The Internet Democracy Project has gone so far as to write that the UDRP was "passed in 1999 without the representation or consent of Internet users" and that it "should be subject to review and vote of reauthorization."

Knowing that such disputes are likely to occur, the unions propose to "reserve the right to identify and resolve possible conflicts prior to completion of domain registrations, in order to prevent disputes arising at a later stage." Presumably this means going out and looking for trouble and trying to resolve it. One wonders if this means going to unions which are not really unions and asking them to please not apply.

The problems facing the sponsoring unions are varied and many and we've barely touched on them here. For example, the ICFTU already raises the question of "segmenting the TLD space" including "devising a system for allocating second and third level names for the purpose of avoiding confusion about which unions are involved in a particular second-level domain, and for directing employees to the union branch or local they may be particularly interested in." They give examples of this -- "iam.boeing.union, seui.boeing.union, iamlocalxxx.union, seiulocalyyy.union". And also this one -- "training.carpenters.union might [which] direct people to carpentry apprenticeships and training links".

They neglect an obvious segmentation which is the national one. What about Ford workers in different countries? You'd need something like "" and "". Or should that be "ford.usa.union" and ""?

This all sounds very exciting from the point of view of cataloging, and I'm sure librarians throughout the trade union movement could be preoccupied with this for years. But how will this make it easier to find trade union websites?

The ICFTU continues its proposal by addressing the question of the expected size of the domain. "It is hard to estimate the number of potential registrations likely to take place in a trade union TLD," they say.

Their database contains "2,000 trade union organisations," which seems to me like an awfully small database. There are already well over 2,000 trade union websites, so this database only scratches the surface. The ICFTU is aware of this, and adds that the "purpose of this database . . . is primarily to maintain information on international and national trade union organisations, and not union locals or branches, many of which already have their own web site."

Their estimates of how many potential users of the unions TLD are quite low. They write: "We expect the overall number of union organisations in the world to be considerably higher [then the 2,000 in their own database] - for example, there are about 5,000 registered unions in India and Bangladesh alone. A recent estimate put the figure at 100,000 trade union organisations world-wide."

I don't know who made that estimate, but they're way off base. There are not 5,000 unions in India and Bangladesh; according to the 1994 Indian Labour Year Book -- the last figures we have -- there are 52,773 unions in India alone. There are something like 50,000 trade union locals in the USA. Taking into account the fact that local, branch unions will eventually have websites -- and the ICFTU proposal assumes this -- the number is in the hundreds of thousands.

The potential constituency for a unions top level domain is much larger than the ICFTU estimates, and thus the need for a much larger structure for handle disputes and appeals is also evident.

One of the more explosive implications of a union top level domain is touched on at the very end of the ICFTU's proposal, buried in a long list of common sense material about various ways in which a union domain would be used -- e.g., some unions like the ICFTU might choose to have icftu.union in addition to The explosive bit reads as follows: "We would also support the use of a service, product or company name in connection with a trade union TLD."

What does this mean in plain English? It means that domain names like microsoft.union and mcdonalds.union and even bigmac.union become possible. The corporations are not going to love this at all. And not surprisingly, they don't. (See in the Appendices, a business view of the proposed union TLD, for more details.)

ICANN is specifically committed to the protection of trademarks (this is know as "intellectual property rights"). To campaigners against virulently anti-union companies, the possibility of creating walmart.union and amazon.union and so on will be awfully tempting.

The very last sentences of the ICFTU proposal indicate an understandable hesitancy about the whole idea: "In order to make a decision which is fully backed by the unions we represent around the world, and in order to fulfil our constitutional obligations, we are bound to consult widely . . . Among the international trade union bodies, and the affiliates of those bodies, we are currently holding an internal discussion on whether or not we proceed in making a proposal to ICANN. There are many issues which need to be examined before we proceed, and we may very well decide either not to proceed, or to defer an application for a TLD to a later date (that is, we may decide that it is too early for us to decide to apply to be one of the first 'test bed' TLDs)."

Alternatives to a union TLD

Are there alternatives to asking ICANN for a union top level domain? Sure there are. But some require rather creative thinking. Offhand, I can think of four alternatives to the proposal for a top level union domain name and I'm sure you can think of others. Here's my list:

And now, in somewhat greater detail:

Do nothing. Stick with the current names and decide that domain names are over-rated. People who are looking for trade union websites will not find them by guessing an address like "boeing.union"; far better to spend the money on promoting existing sites and creating new ones. Using good names in document titles and meta tags ensures better results on search engines anyway.

Make a directory. Create the ultimate list of union websites -- which must be under the control of the unions themselves. There have been several attempts to create global lists of union websites, with the most successful so far having been Steve Davies' Cyber Picket Line. On the LabourStart website, I created a page a couple of years ago called the Global Labour Directory of Directories listing the various sites which were trying to create the ultimate directory for trade union websites. Most of these directories were geographically limited, or limited to particular sectors.

I no longer think that we need to create a central list of links or even a database hosted by the unions. But I do suggest using an existing central resource -- Netscape's Open Directory Project ( -- for this purpose. This is a terrific database of websites maintained by tens of thousands of volunteer, human editors. Unions can begin to take control of their own space in the ODP -- an example can be found today in Britain, where the giant MSF union has its own directory on the ODP, maintained by one of the union's officers. Instead of maintaining a directory of MSF branch websites on the main union website, the directory is maintained on the ODP, using its database features. Among other things, as the ODP is searched by Google and other search engines, this guarantees greater exposure to union websites than a simple list of links buried somewhere on the union's national site.

Try a single domain name. I warn you: this is really going out on a limb, but I wonder if anyone has talked about this seriously. Let's say we all agreed that a joint domain name would be useful for unions. There's a very simple way to do this today without going through ICANN or paying them a cent.

Nothing prevents the sponsoring unions now pushing the union top level domain from simply adopting a single domain name which could be used by unions everywhere. By that I mean they could easily have picked something like and then allowed unions to have subdomains like,,, and so on, or even a simple directory structure -- like or

I don't mean hosting all union websites on a single sever -- I do mean creating a system that allows for easily guessable union website addresses. The single server would then redirect from, say, to the real website, which could stay at

There are some huge advantages to doing things this way. For example, the sponsoring unions, and not ICANN, would have complete and final control over who gets to use the cyber union label which would be created. State-controlled labour fronts and company unions could whine all they want, but just as they cannot contest the ICFTU's ownership of the domain, so they would not be able to contest the ownership of whatever global domain name was picked for the project.

The problem of language would be solved. Instead of worrying about "aliasing" top level domains to different languages, the sponsoring unions could simply register as many domain names as they liked, such as At this level of Internet architecture, aliasing certainly works and is used all the time.

Such a solution is admittedly not ideal. It does little to solve the problem of the cost of moving over thousands of websites email addresses to a new domain. And it would compel unions to route all their web traffic through servers controlled by the ICFTU and its allies, which would presumably be placed strategically around the world. But let's not forget that in accepting the union TLD they were also ceding a bit of their sovereignty to the global trade union bodies.

Its great advantage is that it would create a global cyber union label without dependence on an undemocratic, unelected, unaccountable body like ICANN and all the problems that entails.

Network. There have been several good attempts to link up trade union websites in a global network:

An example is the Union Ring, launched several years ago by Eugene Plawiuk, a school custodian in Edmonton, Canada, whose ever-growing empire of websites began with "Plawiuk Pontificates" and continues with Labournet Canada and many others. Plawiuk was an early adopter of the Web Ring concept, which basically works like this: you set up a Web Ring at the central Web Ring website ( You're given a bit of code (in HTML) which people who choose to affiliate to your website can use on their sites. This code gives visitors to those sites the option to go on to the next site in the Web Ring, or the previous one, or a random one, or even a list of all sites which have affiliated to the ring. It also includes a link to those webmasters who want to know more about the ring and how to join it. Web Rings are useful for surfers, for people with time on their hands who wish to travel from site to site within a particular subject area. They are immensely popular and there are now over 80,000 such rings.

Union Ring took off quickly and has today reached the extraordinary number of around 650 affiliated labour websites. What is remarkable about the project is that many large unions, often quite wary about what they will link to, have not hesitated to put Union Ring code on their front pages. It has become a kind of identifying badge, a "union label" as it were, for hundreds of trade union websites.

Plawiuk has done little to promote the ring. What seems to drive trade union webmasters to sign up is its ubiquitous appearance all over the net. And because of its truly global character, displaying a Union Ring logo on one's website is the twenty-first century equivalent of hoisting a red flag over a trade union building in the nineteenth or twentieth centuries. It has become a globally recognised icon of the labour movement, a unifying trademark made possible only by the net.

In the spring of 1999, seeing that Union Ring was getting unwieldy with its hundreds of affiliates, LabourStart initiated a UK Labour Webring which has managed to attract more than thirty British trade union sites.

A much less successful effort than the web rings has been the attempt to create a banner exchange network linking trade union websites across borders. An early effort, begun in 1997, used the Link Exchange system, now owned by Microsoft. Link Exchanges work in a similar way to Web Rings, but require more work on the part of webmasters, who must design banner advertisements that are pooled and shown across the network. They usually work on a 2:1 ratio, meaning that for every two times a Link Exchange banner ad is shown on your site, one of your ads is shown on another site in the network.

Link Exchange accepted LabourStart as a partner in its affiliate program and at my suggestion created a special category for trade union websites that had joined the program. This was promoted through LabourStart under the name "Labour Link Exchange", but after more than two years, only about 50 sites had signed up. Furthermore, though a site could choose to target its banner advertisement at other trade union websites, it had little control over whose ads appeared on its own site. In the Union Ring, the only sites one was linking to were other labour sites -- but not in Link Exchange.

A more promising effort currently underway is the new International Labour Banner Exchange (ILBE), launched in January 2000. Unlike the original Link Exchange, the International Labour Banner Exchange rotates advertisements only among those websites which join it -- making it a kind of mini-Link Exchange open only to trade unions. As I write, it already has nearly 40 websites signed up, though this is admittedly only a tiny percentage of the total number of labour websites.

Webrings and banner exchanges have opened up new possibilities for trade unions in different countries to assist one another by driving traffic to each other's websites. In addition to that, and like the highly successful Union Ring, they provide labour websites with a way of proudly declaring their affiliation with a global network. These are proven and tested ways of doing one of the things the proposed union top level domain aimed at.

There are even more exciting new ways unions can work together using the new technology that would have been inconceivable only a short time ago. One of these is the Labour NewsWire (LNW), launched in March 2000 by LabourStart. LNW is basically a small computer program written in the Perl scripting language, which resides on LabourStart's web server in the UK. It is linked to both a JavaScript program and a Netscape RDF (Resource Description Framework) file. The technique is impressive and all the work was done by two volunteers, Bruno Postle and Chris Croome of WebArchitects in Sheffield. What the script does is allow any website in the world to include labour news headlines from LabourStart -- which are updated 24 hours a day, seven days a week -- on their site. An early version of the LNW required users to install and configure a complex CGI script on their web servers, and this proved to be an insurmountable obstacle for many. The new version required only that webmasters copy and paste a single line of code from the LabourStart website to their pages.

LNW was immediately adopted by the Congress of South African Trade Unions (COSATU) which integrated lists of both global and South African labour news stories in their site. COSATU was swiftly followed by the main website of the Russian trade unions,, which neatly incorporated the latest worldwide labour news headlines in a box on its front page. (The news headlines are in English, as are the news stories they link to, but the rest of the front page is entirely in Russian.) Within a few weeks, LNW was adopted by several dozen more websites, including such high traffic sites as the New South Wales Teachers Federation in Australia and Eugene Plawiuk's Labournet Canada.

The closest parallel to a pre-Internet version of Labour NewsWire would have been a wire service, like Associated Press or Reuters, but one owned by and serving the labour movement. Theoretically, such as service could have been created decades ago, possibly at the instigation of (or with the approval of) the ICFTU itself. It could have been providing a continuous news feed from unions around the world to the labour press.

For trade union webmasters, the incorporation of Labour NewsWire in their sites is a way to always and automatically have current information on the site. For the international labour movement, it is potentially a consciousness-building tool that will teach trade unionists everywhere to think globally.

The Union Ring, the International Labour Banner Exchange, and Labour NewsWire demonstrate ways in which trade union websites can work together, and indeed label themselves as labour sites, without recourse to a common top level domain name.

As I was writing these words, I put up one of those online public opinion polls on the LabourStart website and asked people if they thought having a union top-level domain sounded like a good idea. Of the several hundred people who replied, the overwhelming majority -- well over 80% -- said yes.

I think that trade unionists are often eager to use the new technology, to give unions a much higher profile in cyberspace. And they're right about that.

But as I hope I've made clear above, there are a lot of problems with the current proposal for a union top level domain and until these are resolved, it would be a mistake for the labour movement to rush into this.

Other issues

The question of a union top level domain is an interesting one, but I don't want to create the impression that this is the only, or even the most important, question facing the bodies governing cyberspace today.


One of the biggest issues on the agenda for the next period is the internationalization of the Internet, which actually sounds a bit funny, being that we all talk about the "world-wide" web. When Tim Berners-Lee invented the web more than a decade ago, the technology only worked well with the English language. The earliest versions of HTML were "seriously restricted by its reliance on the ISO-8859-1 coded character set, which is appropriate only for Western European languages" -- according to one of the landmark documents published by the Internet Engineering Task Force's HTML Working Group, known as RFC 2070. This document, which was apparently first drafted in 1994, sought to expand the possibilities of coding web pages in languages other than English for the first time.

I should emphasize that even the document's mentioning of "Western European languages" was a bit far fetched -- one simply could not include anything but the standard 26 English letters without any accents in HTML.

One might point out that 1994 was really the very beginning of the web, and was a very long time ago, and what does this have to do with us today. But the point is that the standards bodies that set up the web -- first the IETF HTML Working Group, today the World Wide Web Consortium -- were set up by researchers, academics and businessmen without any input from civil society. There was simply no one around from, say, China, to point out that if HTML relied on the English language character set it would have somewhat limited value for the vast majority of countries and people.

The internationalization of HTML, and particularly the development of Unicode, which is the best way so far to truly internationalize the web, grew out of the desire by corporations to start using the web commercially. If there's a profit in it, the bodies governing cyberspace will make the appropriate changes. This may sound perfectly reasonable to some ears, but this is not the only way things can be.

Compare how the IETF, W3C and others were thinking to how more traditional international institutions -- the ones that have some input from civil society -- work.

The United Nations, the European Union, the global trade union bodies all work in languages other than English. In 1994, when the possibility of using HTML in languages other than English was first raised, no one had to suggest to any of these bodies that they take into account languages other than English in their daily work. They had done so from the beginning.

Even today, I'd argue that the W3C doesn't go nearly far enough in internationalizing the web. Let me give an example.

As I began training trade unionists to set up and maintain their own websites here in the United Kingdom, I began to hear things like, "Oh, do we have to spell 'centre' in the American way?" And it struck me that there is no reason why programmers coding in HyperText Markup Language must be compelled to use the American spelling for everything. Certainly it would not be difficult for the W3C in some future version of HTML to allow alternate spellings in English, allowing "centre" as a synonym for "center".

Then I attended an international conference in London about the new economy. One speaker got up, nervously addressing the crowd in a language that was not his own -- English -- and began by telling us that every single computing language which we are talking about, the building blocks of the new economy, such as HTML, Perl, C++, Java, PHP and so on -- every single one of these without exception is based on the English language. (And, one might add, American English at that.)

These languages are known as high level languages because they are relatively easy to learn (if you know English) as compared to machine language, which is essentially numeric code. I learned about these things in my early days as a programmer in the mid 1980s. I began writing my first code in a language called RPG (which stood, originally, for Report Program Generator) and this language used English words like "add", "if", "end" and shortenings of words like "begsr" for "begin subroutine". And all this made programming a lot easier than having to write out the code in the language which the machines themselves understood -- machine language.

Programming languages are, in essence, a way of communicating between humans and machines. At one end, the machines understand only two things, zero and one. Or to put it another way, on or off. At the other end, humans understand, each one of them, at least one of the 6,000 or so languages spoken on the planet today. Programming languages are supposed to be easier when they more closely approximate human languages than machine languages. So I guess compared to what programmers were going through in the 1950s and 1960s, having to talk to computers in their own binary tongues, languages like RPG and much later, HTML, were huge improvements.

But they were still only in English. And being that English has been the language of all major programming languages for a generation, it is not surprising that the authors of so many of the famous ones are native English speakers. HTML, for example, is the work of the Englishman Tim Berners Lee. Java was invented by American programmers working for Sun. JavaScript was created by Netscape programmers.

Even when people who are not Americans get into the act, they are compelled by the nature of the digital economy to do their work in English. Thus the Finn Linus Torvald, creator of Linux, that extraordinary variant of the Unix operating system that can run on an ordinary PC, wrote up the code in English. Had he chosen to write it up in Finnish, it would probably not be the serious challenger to Microsoft that it is today.

There is no technical reason why this should be the case. For example, in HTML, an ordinary web page is broken up into two parts, the header and the body. These are delimited by tags that look like this:

<head> </head>
<body> </body>
HTML files with these kinds of tags sit on a server somewhere on the net and are accessed by client programs known as browsers, such as Microsoft Internet Explorer, Netscape Navigator and Opera. Nothing prevents those client browser programs from being able to understand alternative terms for header and body in other languages. In fact, those client browsers are already customized to work in non-English languages and properly interpret the uncoded bits that go between the tags, which can be written in any language.

So for example, if I were to design a web page in the Esperanto language, I know that anything I write between the <body> tag and the </body> tag in Esperanto characters will be correctly interpreted by the client browser program -- if that has been configured to understand Esperanto (allowing the browser to display those few letters which are not to be found in English).

Common sense tells me that if the client program (the browser) can be configured to correctly render all the Esperanto letters, why can't we write the HTML code itself in Esperanto? Instead of the <head> tag, why can't I use <kapo> (which is Esperanto for head) and for the <body> tag, <korpo> (obviously, Esperanto for body).

Then the structure of an Esperanto language document could be something like this:


And the same would be case for any other language, including those not written in Latin characters.

How would browser software know how to interpret tags in non-English languages? Simple -- build this into the <html> tag which should head up every HTML document on the web. This is what we are supposed to be doing anyway when we write up a web page. (Though when was the last time you added a line to a web page telling any browser reading it that this was written in American English? Probably never, but programmers preparing web pages in Mandarin, Arabic and Hindi certainly are doing this all the time.)

Critics will say that there is no need for all this because programmers know English anyway. Really? Has anyone visited a bookshop in Berlin or Brussels lately? I have. There are limitless books on website design, explaining how to use HTML code to create web pages -- but the books are written entirely in German, Flemish and French. And this is certainly true of bookshops in Buenos Aires, Beijing, and Be'er Sheva. As the population online soars from its present 330 million to a billion and more, with millions and millions of people creating their first websites, why is it necessary that they all work in American English to do so when the content of their websites may be written in Hindi, Mandarin or Korean?

The reason why is that the body responsible for HTML, for maintaining a single global standard for web sites -- which is important, I admit -- is based in the United States, headed up by an Englishman, and working in the old fashioned way with little or no input from civil society. I should correct that -- there is some input from outside the community of techies that normally discusses these issues. One component of civil society does get invited to the table these days to discuss these kinds of things. That component is the business community.

The group I'm talking about, the World Wide Web Consortium, needs the involvement of representatives of civil society, and most importantly, the trade union movement, to ensure that concepts like "internationalization" do not only mean that Coca Cola can get its website read by potential customers in Indonesia, but also that Indonesian trade unions and other popular organizations, can get their messages out in their own language -- and that the programming language they will need to learn to do so will be as user-friendly as possible. User-friendly in this case means that they will not have to learn the meanings of "body", "head" and so on.

The possibility of even further internationalization was suggested in the ICFTU proposal for a union top level domain. They suggested that there be "aliases" in different languages for the top level domain "union".

But why stop at the union TLD? Instead of the handful of global top level domains which exist today and are likely to exist in the near future, why not allow the translation of all global top level domains into -- at the very least -- the most widely used languages in the world?

Today it's nearly impossible to create website addresses with non-English characters, though some have tried. But there is no technical reason why this should be the case.

In fact, just a couple of weeks ago, Network Solutions, Inc., which until a short while ago had a monopoly on registration of the most popular domains (com and org), announced that it was launching a test program to allow the registration and use of web addresses in 55 languages including Japanese, Chinese, Korean, Arabic and Hebrew. This is good news -- but why did it take until the year 2000 for it to happen?

The international trade union movement, with its historic commitment to multilingualism and respect for other cultures and languages, is uniquely qualified to play a role here. Bodies like the United Nations, World Bank and WTO will choose to use certain languages based on a nation's perceived economic or military power. While those bodies have their own criteria for resolving language matters, the unions have to work differently.

The involvement of international trade union representatives in the decision making process around the future of Internet programming languages -- including not only HTML but Cascading Style Sheets (CSS), Extensible Markup Language (XML), Wireless Markup Language (WML) for mobile phones, and so on -- will bring a breath of fresh air into those discussions. That is one reason why trade unionists and other representatives of civil society need to be involved not only in ICANN, but in the deliberations of the W3C, ISOC, the IETF, and so on.

I can imagine objections to this coming from both the techies and the business leaders who currently run these things, but they can be answered by the arguments above -- and by the demonstrated technical competence of many in the trade union movement itself. There are increasing numbers of labour websites that reveal a proficiency in and understanding of the new media, and the people involved in those sites can certainly play a useful role in the W3C.

Trade unionists themselves might object, saying that we have more important things to do than be involved in the decision making process around a new version of Cascading Style Sheets. And they're right. Given a choice between running a campaign to get jailed trade union leaders released and attending a symposium on XML, the former is indeed more urgent.

Nevertheless, for the first time in history, we do have a chance to have a say in technology we are all going to be using. Decisions are being made which will affect us. We cannot stand aloof and allow others to make these decisions for us, particularly when the main concern of those others is to increase their profits.

Intellectual Property Rights

I've started with the question of language, but there are other issues as well that should concern the trade unions.

For example, ICANN has shown a tremendous sensitivity to the question of intellectual property rights when it comes to domain names. I should explain what the problem is here.

Domain names such as were not automatically given out to corporations that own the trademarks. Nothing would have prevented individuals, organizations, or even rival corporations from spending a paltry sum and buying up such a name. Many people earn a living this way. They are known as cyber-squatters and they buy up domain names that they think will sometimes have monetary value.

Some of these names are generic -- such as -- and some are the names of actual corporations. Often the problem is solved when the corporation simply buys up the name. This happened to Salon, an online magazine, which found that its logical domain name -- -- was held by, if I remember correctly, a hair styling salon. After a few years, Salon bought the rights to the domain name.

Sometimes the purchase of a particular domain name can be seen to be malicious. For example, visit and you'll find a pornographic website. The real address for the White House is Many sites have tried to make money by adopting addresses that are close to those of well-known, heavily trafficked sites -- such as,, and so on. (I know, this seems a bit pathetic, but this is how some people earn a living.)

In some well-publicized cases, celebrities have gone to court to get back the rights to their "own" names that have been registered by cyber-squatters.

And ICANN, alert to the issues, has recognized the important of dispute resolution based on the principle of intellectual property rights. In other words, Nike has a kind of "natural right" to the domain name After all, it's their trademark. They invented it, they registered it, and it has value to them as a business.

What does this have to do with unions and the rest of civil society? It means that ICANN will line up on the side of Nike when and if some tries to register a domain like nike.humanrights (humanrights is one of the proposed new top level domains). If the domain nike.humanrights is registered by one of the groups campaigning against Nike's exploitation of child labour around the world, its denial of trade union rights, and so on, you can bet that Nike will rush to the dispute resolution process screaming about its intellectual property rights. And it does this not to prevent a competitor with using its name on competing products (which was, I guess, the whole point of trademarks) but to prevent people from visiting a site with a catchy URL like nike.humanrights.

An employers organization in the US has already written to ICANN suggesting that the whole idea of a union TLD be dropped (see Appendices for the full text) and though they do not explicitly address the intellectual property rights issue, they do say that "by having a company.union TLD, employees may be led erroneously to believe that they were already represented by a union that was recognized by a company." In other words, companies will want to have a say about how "their" trademarked name is used, even when the user is an independent trade union.

Trade unions will not have a simple, black-and-white position on intellectual property rights. Many members of trade unions -- journalists, musicians, software programmers, and so on -- have a stake themselves in copyrights. Unions too often feel that they must protect their own "brands" (yes, unfortunately, some trade unionists do speak of "brands").

Some on the left are completely dismissive of the notion of intellectual property rights, though few have suggested a way to compensate people who earn their living (and that doesn't only mean giant corporations) from the enforcement of such rights. One solution would be to abolish capitalism and create a new society based on the principle of "from each according to his or her ability, to each according to his or her need." But that's the subject for another book, isn't it?

Control of the domain name system

In addition to internationalization and intellectual property rights, there are other issues that fall under the sway of ICANN, ISOC and the W3C that should concern unions.

For example, there was the recent (and as yet unconcluded) battle between ICANN and some of the national registrars for country code top level domains (ccTLDs). After ICANN received control of the domain name system from the US government, it did what comes naturally to any privatized company taking over part of what was formerly public sector: it began sending out bills.

ICANN came up with huge fees for the national registrars, some of which rebelled and refused to pay, such as the European registrars. Others, such as the South Africans, claimed an inability to pay the thousands of dollars ICANN was demanding. There is a serious danger that ICANN may one day enforce its control by blocking out a domain that has not paid its bills.

The issue of ICANN's fees is linked in my mind to the question of who actually controls the "root servers" which determine the flow of traffic -- both on the web and email -- through the Internet. There are a dozen of these, three outside the US (in Stockholm, London and Tokyo) and all the rest are inside the USA. Those root servers are owned today by the US government, not ICANN.

Knowing that to be the case, I wouldn't be surprised to see the occasional "rogue state" blocked off from the rest of the Internet. Some readers may find it hard to imagine the US government and ICANN deciding to invalidate the Yugoslavian ccTLD, for example. That really wouldn't be playing fair. But whatever one thinks of NATO's war last year against Serbia, playing fair was hardly a consideration, as the bombing of Yugoslavian television demonstrated. If Washington was prepared to do that, killing journalists and other civilians, why would they think twice about shutting down a domain on the Internet if it served the US's national interest? Of course they would do it -- and not only Yugoslavia's domain.

Shutting down a domain name means that websites bearing that name will be inaccessible, because we are all dependent upon the handful of existing root servers. In addition to that, we will not be able to send email to addresses bearing barred domain names. In an age of electronic warfare, shutting down an enemy's domain on the net is the very least we can expect the military to do.

On this kind of issue, the position taken by the labour movement should be quite clear: we should demand that the root servers be entirely removed from the control of the US government and that ICANN be banned from ever shutting a country off from the Internet, regardless of whether or not it pays its bills. (In the latter case, the dispute should be seen as one between ICANN and the local registrar; the whole country should not be punished by being barred from the net.)

There is now some reason to believe that the decades-old domain name system over which ICANN presides may be on its last legs. Opponents of censorship and centralized control of the net have been developing alternative ways to route traffic, such as Freenet. Campaigners against child pornography and others have expressed concern that these technologies will make it much harder or even impossible to track down and then shut down offensive websites -- which is, of course, the whole point. As these technologies spread, ICANN may become increasingly irrelevant.

There are many more issues being discussed and many more will arise in the years to come. It is hard to imagine most trade unionists today even thinking about some of these questions. Even getting a discussion going of the proposed union TLD has been difficult. But these are important questions.

We must be involved in the discussion and we must have a voice in the decision making process. Because the Internet belongs to us too.


In this short book, I've tried to raise some of the issues that should concern trade unionists regarding the governance of cyberspace. There are other issues and there is more to discuss; this is only the beginning.

Nevertheless, the outlines of a possible labour program have emerged from these pages. So to conclude, I'd like to suggest the following five point program:


Acronyms Used

ccTLD country code Top Level Domain (such as .uk, .fr, .de)

COSATU Congress of South African Trade Unions

CSS Cascading Style Sheets

DNS Domain Name System

DNSO Domain Name Supporting Organization

gTLD global Top Level Domain (such as .com, .net, .edu)

HTML Hyper Text Markup Language

HTTP Hyper Text Transfer Protocol

ICANN Internet Corporation for Assigned Names and Numbers

ICFTU International Confederation of Free Trade Unions

IETF Internet Engineering Task Force

ILO International Labour Organization (also: International Labour Office)

IMF International Monetary Fund

ISO International Standards Organization

ISOC The Internet Society

ITS International Trade Secretariat

LNW Labour NewsWire

OECD Organization for Economic Cooperation and Development

PHP Personal Home Page

RFC Request for Comments

RPG Report Program Generator

UAW United Auto Workers

UDRP Uniform Domain Name Dispute Resolution Policy

UNI Union Network International

URI Uniform Resource Identifier

URL Uniform Resource Locator

W3C World Wide Web Consortium

WFTU World Federation of Trade Unions

WIPO World Intellectual Property Organization

WML Wireless Markup Language

WTO World Trade Organization

XML Extensible Markup Language


1965 - The US Department of Defense's Advanced Research Project Association begins work on ARPANET, forerunner of the Internet.

1969 - Researchers at four US campuses create the first hosts of the ARPANET, connecting Stanford Research Institute, UCLA, UC Santa Barbara, and the University of Utah.

1971 - The ARPANET grows to 23 hosts connecting universities and government research centers around the USA.

1972 - The InterNetworking Working Group becomes the first of several standards-setting entities to govern the growing network. Vinton Cerf is elected its first chairman. Meanwhile, trade union leader Charles "Chip" Levinson publishes his book, International Trade Unionism, with its prophecy of a global labour network and online distance trade union learning.

1973 - The ARPANET goes international with connections to University College in London, England and the Royal Radar Establishment in Norway.

1979 - Tom Truscott and Jim Ellis, two grad students at Duke University, and Steve Bellovin at the University of North Carolina establish the first USENET newsgroups.

1981 - ARPANET has 213 hosts. A new host is added approximately once every 20 days. In western Canada, Larry Kuehn and Arnie Myers of the British Columbia Teachers Federation, create the first online trade union network, linking up all the union's officers with modems and mobile computers. Meanwhile, in Norway, Kristen Nygaard, a computer specialist, proposes a global network to link up trade unions. It is written up in some international labour magazines but proves to be too far ahead of its time.

1982 - The term "Internet" is used for the first time.

1983 - TCP/IP becomes the universal language of the Internet. In Britain, activists at International Labour Reports get their first modem and begin the first, primitive exchanges of messages with the Netherlands; they go on to found Poptel, the Popular Telematics Project, which does much to link up the first trade unions to go online.

1984 - William Gibson coins the term "cyberspace" in his novel "Neuromancer." The number of Internet hosts exceeds 1,000. In Brussels, the International Chemicals Workers Federation -- formerly headed up by Chip Levinson -- begins using email and spreading the technology to its affiliates around the globe. It also pioneers the use of online databases in its research. The London-based International Transport Workers Federation also adopts the new technology.

1986 - Case Western Reserve University in Cleveland, Ohio creates the first "Freenet" for the Society for Public Access Computing. Solinet, the "Solidarity Network", is launched at the initiative of Marc Belanger from Canada's largest union, the Canadian Union of Public Employees. In South Africa, Worknet is launched, paving the way for extensive use of computer communications by unions and others in civil society.

1987 - The number of Internet hosts exceeds 10,000.

1988 - 150 trade unionists from across Canada participate in Solinet's two-month long online conference about technological change. In Denmark, thirty trade unionists are connected in an experiment in online communications called FLOKs.

1989 - The number of Internet hosts exceeds 100,000. A second Danish trade union experiment in online communications is conducted, this time linking up 70 labour activists. It is sponsored by the country's national trade union center, LO. At the time of the Tiananmen square massacre, the embattled independent trade unions in China use email and fax to get their message out to the world.

1990 - According to one study, there are now 180 trade union organizations online around the world. At a conference in the Netherlands, trade unionists and other activists using the net in different countries meet for the first time. Danish, Swedish and British trade unionists conduct an experiment in online distance learning. In the USA, ARPANET is decommissioned, leaving only the vast network-of-networks called the Internet. The number of hosts exceeds 300,000.

1991 - Tim Berners-Lee, working at CERN in Switzerland, posts the first computer code of the World Wide Web in a relatively innocuous newsgroup, "alt.hypertext." The National Science Foundation lifts the restriction on commercial use, clearing the way for the age of electronic commerce.

1992 - The Internet Society is founded. The first global trade union conference about online communications in held in Manchester. Ninety participants from two dozen countries attend. In the USA, the AFL-CIO launches LaborNet on CompuServe. More than 1,000,000 hosts are part of the Internet.

1993 - Mosaic, the first graphics-based Web browser, becomes available. Traffic on the Internet expands at a 341,634% annual growth rate. The quarterly magazine Workers Education devotes an entire issue to the subject of "Computer Communications and the Labour Movement".

1994 - The World Wide Web Consortium is founded. The International Federation of Workers Education Associations becomes the first global trade union body with a website. Though the group is based in Tel Aviv, its website is hosted in California and later is moved to Britain, then Finland. Workers Education begins a regular column called "Online" which reports on unions and worker education groups which launch websites, one by one. The International Labour Organization and the International Confederation of Free Trade Unions co-sponsor a report on the "opportunity and challenge of telematics". The San Francisco Free Press -- the first online daily strike newspaper -- is launched; the strike is won after only 11 days.

1995 - Sun releases the Java programming language. In New South Wales, Australia, unions launch LaborNet; the site is eventually to host the weekly online magazine, Workers Online. Hoping to emulate the success of the San Francisco strikers, locked-out Detroit journalists launch their own strike paper online, The Detroit Journal. It is still publishing in 2000 as the lockout continues.

1996 - Nearly 10 million hosts online. Approximately 40 million people are connected to the Internet. More than $1 billion per year changes hands at Internet shopping malls. The Labour Movement and the Internet: The New Internationalism is published; the accompanying website later evolves into LabourStart. The book lists some 60 labour websites. The ICFTU finally launches its website; it appears in four languages. The first cyber-campaign organized by international unions targets Bridgestone/Firestone with email protests.

1997 - The Labour Webmasters Forum is launched online. South Korean unions and their allies make extensive use of the net during their general strike; later in the year, they host an international conference involving some pioneers of labour use of the net.

1998 - LabourStart is launched as the first portal website for the trade union movement with daily news content. ICANN is founded as the US government waives its right to control the domain name system for the Internet

. 1999 - LabourStart moves over to a database model with remote access, leading to the creation of a global network of volunteer correspondents who keep the site up to date 24 hours a day. Art Shostak's Cyber Unions is published. There are over 2,000 labour websites.

2000 - ICANN holds the first elections in cyberspace.

Useful website addresses

The bodies governing cyberspace:

The Internet Society -

The Internet Engineering Task Force -

The World Wide Web Consortium -

The Internet Corporation for Assigned Names and Numbers -

Critical civil society organizations:

ICANNWatch -

Internet Democracy Project -

Computer Professionals for Social Responsibility -

Trade unions:

LabourStart -

International Confederation of Free Trade Unions -

Education International -

International Federation of Building and Wood Workers -

International Federation of Chemical, Energy, Mine and General Workers' Union -

International Federation of Journalists -

International Metalworkers' Federation -

International Textile, Garment and Leather Workers' Federation -

International Transport Workers' Federation -

International Union of Food, Agricultural, Hotel, Restaurant, Catering, Tobacco and Allied Workers' Association -

Public Services International -

Union Network International -

Trade Union Advisory Committee to the OECD -

Global Labour Directory of Directories -

Sources of statistics:

NUA Internet Surveys -

International Labour Organization -

Governing Cyberspace: "Where is James Madison when we need him?"
by David Post

In a column last Fall I suggested that the pending reorganization of the Internet's domain name system (DNS) had the potential to become cyberspace's own "constitutional moment," a profound and thorough transformation of the institutions and processes responsible for law-making and regulation on the global electronic network. Over the last several months, the shadowy outlines of a new kind of constitutional structure for cyberspace have indeed begun to emerge. The consequences of these developments for the Internet's future could not be more profound, and the picture that is emerging is not always a pretty one. Not many people are paying close attention to these developments; they should. Not to be an alarmist, but to boil a frog alive, the parable tells us, you need just to turn up the heat by increments so small that the frog never notices-never pays attention to-the rising temperature until it is too late to do anything about it (like jump out of the pot). Well, the temperature on the Internet is starting to rise.

(A bit of) Background

[Those of you who already know how the DNS works, and what a "root server" is, might want to skip ahead . . . ]

For the Internet to exist as a single coherent network, there must be a way to be sure that a message sent from any point on the network to, or a request to view the webpage at, is routed to the "right" machine; that is, each of those addresses ( or must be associated unambiguously with a particular machine if message traffic is to move in a predictable way.

To make an extremely long story short, the global network we call "the Internet" manages this by, first, requiring that each machine on the network have a unique numerical address [e.g.,]; indeed, to be "on the Internet" means to have (or to be connected to) a machine to which such an address has been assigned. For your message to to be routed correctly, your computer must somehow be able to find the numerical address corresponding to the machine named This is accomplished, on the Internet, by means of what Tony Rutkowski calls a "magical mystery tour."

When you send off a message requesting a copy of the home page (""), the message first stops off at a machine known as a "DNS [Domain Name System] server." It is the job of the DNS server, which is usually operated by your Internet Service Provider, to find the correct numerical address for your message. The DNS server reads, in effect, from right to left; seeing that this is a message destined for some machine in the EDU domain, it needs to find out where addresses in the EDU domain are stored. It does this by asking a different machine (known as the "root server") that very question: "Who is responsible for the EDU domain?" The root server replies with the numerical address of a different machine (known as the "EDU domain server"). Your DNS server then asks the EDU domain server the question: "Who is responsible for "" The EDU domain server replies with another numerical address [or with "Not Found" if it cannot find an entry for "" in its database of names and addresses]. Your DNS server then asks this machine: "Who is responsible for" replies with yet another numerical address, and now your DNS server has completed its task; once it receives the address for, it places that address into your message and sends it on its way.

How does this all work as smoothly as it does? Who is in charge of the root server? How does the operator of the root server decide which machines are the "authorized" domain servers for EDU, or COM, or ORG, or any of the other top-level domains? Who controls those machines (and the database of names and addresses contained in them)? And how is this whole scheme enforced? That is, what makes the root server "the" root server? Why do the many thousands of Internet Service Providers, operating the many thousands of DNS servers worldwide, all use the same root server? In the early days of the Internet, of course-through, say, the early 90s-no one outside the small cadre of engineers that was putting the system together cared very much about the answers to these questions. There were, of course, answers to them all. The United States government had long operated the root server (a holdover from the days that the Internet was a Defense Department project), and had worked with something known as the Internet Assigned Numbering Authority (IANA) -- a group of engineers led by the late Jon Postel-to organize the necessary data and to see that the various domain servers were being properly managed.

As long as it all seemed to be working smoothly enough; who cared what was going on behind the Wizard's curtain? And who noticed when, in 1992, as the extraordinary growth of the network began to outstrip the management capacity of this (largely volunteer) operation, the U.S. government engaged a private firm, Network Solutions, Inc. (NSI), to manage and maintain the databases and domain servers for the COM, ORG, and NET domains?

But slowly, as more and more people began to realize that the Internet was a Really Big Deal (and that these funny "domain name" things might actually be of real value), more and more people started to pay attention to all of this, and this arrangement began to come under increasing fire from many quarters. The government and NSI found themselves increasingly under attack from within and without the Net community by those challenging NSI's apparent monopoly control over these increasingly valuable top-level domains, by trademark owners concerned about domain names that appeared to infringe upon their valuable trademark rights, and others.

As the expiration date of the government's contract with NSI approached last June, the Commerce Department announced that the government wanted to get out of the DNS management business entirely. Citing "widespread dissatisfaction about the absence of competition in domain name registration," and the need for a "more formal and robust management structure," the government proposed transferring responsibility for management and operation of the DNS to a private non-profit corporation. This new corporation, to be formed by "Internet stakeholders" on a global basis, would take over responsibility for overseeing the operation of the authoritative Internet root server, and would be charged with introducing competitive, market mechanisms into the allocation of Internet names and addresses.

Um, What Does This Have To Do With "Internet Governance"?

This new corporation-ICANN, the Internet Corporation for Assigned Names and Numbers-has, over the past several months, set up shop and gotten to work. It's been a busy time. It has begun to establish "Supporting Organizations," new coalitions comprising various Internet constituencies (e.g., domain name registrars, trademark owners) who will be responsible for electing members of the ICANN Board of Directors and for formulating aspects of domain name policy. It has taken the first steps towards introducing competition into the domain name system, accrediting five companies (America Online, CORE (Internet Council of Registrars), France Telecom/Oléane, Melbourne IT, and to begin issuing registrations in the COM, ORG, and NET, domains during a two-month test period (along with twenty-nine other entities who can begin accepting registrations in these domains once the test phase is completed). It commissioned, and recently adopted (in part), a report from the World Intellectual Property Organization (WIPO) outlining the procedures to be used in cases involving "cybersquatting" (the intentional "warehousing" of domain names for later sale).

My goal here is not to discuss any of these specific actions; there is much here to digest and debate, pro and con, and I will have a great deal more to say about the specifics of ICANN's activities over the next several months. [As, I hope, will others; I particularly recommend Michael Froomkin's commentary on the WIPO Report referenced in the preceding paragraph, posted at his website.]

Rather, my goal here is just to suggest that notwithstanding the government's (and ICANN's) protestations to the contrary, this is about nothing less than Internet governance writ large. The Commerce Department took pains to characterize it in other terms; this new corporation would be responsible only for "technical management of the DNS," the "narrow issues of management and administration of Internet names and numbers on an ongoing basis"-sort of what the International Telecommunications Union does with respect to managing interconnections on the international telephone network. This new framework for managing the DNS ". . . does not set out a system of Internet "governance." Existing human rights and free speech protections will not be disturbed and, therefore, need not be specifically included in the core principles for DNS management. In addition, this policy is not intended to displace other legal regimes (international law, competition law, tax law and principles of international taxation, intellectual property law, etc.) that may already apply. The continued applicability of these systems as well as the principle of representation should ensure that DNS management proceeds in the interest of the Internet community as a whole."

It is all well and good to say that this new institution will not be engaged in Internet governance-but words will not make it so. Any entity exercising control over the DNS will be subject to immense pressure to do more than mere "technical management," because, bizarre as it may seem at first glance, the root server, and the various domain servers to which it points, constitute the very heart of the Internet, the Archimedean point on which this vast global network balances.

To appreciate that, imagine for the moment that you had control over operation of the root server. You alone get to decide which machines are "authoritative" domain servers for the COM, NET, ORG, EDU, and the other top-level domains, the machines to which all Internet users worldwide will be directed when they try to send any message to any address in those domains. You have the power, therefore, to determine who gets an address in those domains-who gets a passport without which passage across the border into cyberspace is impossible. You can say "From now on, we will use the data in machine X as the authoritative list of COM names and addresses, but only so long as the operator of that machine complies with the following conditions," and then you can list-well, just about anything you'd like, I suppose. It's your root server, after all. You can require that all domain server operators pay you a certain fee, or provide you with particular kinds of information about the people to whom they have handed out specific names and addresses, or only allow transmission of files is a specified format, or abide by a particular set of laws or rules or regulations. And you can demand that they "flow through" these conditions (or others) to anyone whom they list in their authoritative databases, that they revoke any name given to anyone who does not pay the required fee, or provide the required information, or use the specified file format, or comply with the specified rules and regulations.

This is quite literally a kind of life-or-death power over the global network itself, because presence in (or absence from) this chain of interlocking servers and databases is a matter of [network] life or death: If your name and address cannot be found on the "authoritative" server, you simply do not exist-at least, not on the Internet. Eliminate the entry for from the COM domain server and vanishes entirely from cyberspace; designate as the new COM domain server a machine that does not have an entry for in its database, and you have imposed the electronic equivalent of the death penalty on

Anyone interested in controlling the rules under which activities on the Internet take place-and many commercial interests, who now realize the huge economic stake they have in this medium, and many governments, who have spent the last few years worrying about how they would ever get back their taxing and regulatory authority over Internet transactions, find that they are indeed quite interested in that now-is likely to find the existing of a single controlling point awfully tempting. Anyone with a vision of how the Internet can be made a "better" place " by making it safer for the exploitation of intellectual property rights, say, or by eliminating the capability to engage in anonymous transactions, or by making it more difficult for children to get access to indecent material-is likely to view control over the root server as the means to impose its particular vision on Internet users worldwide. After all the talk over the past few years about how difficult it will be to regulate conduct on the Internet, the domain name system looks like the Holy Grail, the one place where enforceable Internet policy can be promulgated without any of the messy enforcement and jurisdictional problems that bedevil ordinary law-making exercises on the Net.

And that is why these are governance questions, why any reorganization of this system, far from being an arcane technical detail of Internet engineering, is inherently of constitutional significance. Power corrupts, absolute power corrupts absolutely-on the Internet as elsewhere. Questions about constraining any form of absolute power are constitutional questions of the highest order, and "governance" means nothing more (and nothing less) than the search for mechanisms to insure that absolute power is not exercised in an unjust or oppressive manner. How can we be assured that ICANN will be able to resist pressures to stray beyond this limited "technical" mandate? Where are the checks on the new corporation's exercise of its powers?

You think, perhaps, that I exaggerate the significance of these developments, and perhaps I do. But let me point to a few dark clouds on the horizon that make me very, very nervous about what ICANN is up to. Remember all those things you could do if you were in control of the root? Like " require that domain server operators pay you a certain fee"? Well, ICANN has imposed the requirement that each accredited registrar pay ICANN a fee of $1 for each new domain name they hand out-can anyone say "taxation without representation"? Or "provide you with particular kinds of information about the people to whom they have handed out specific names and addresses"? ICANN, having now adopted the WIPO Report referenced earlier, is about to impose a requirement on all domain name registrars that they collect and make available "accurate and reliable contact details of domain name holders," and that they agree to "cancel the domain name registrations" wherever those contact details are shown to be "inaccurate and unreliable."-a move with grave consequences for the continued viability of anonymous communications on the Internet. Or "abide by a particular set of laws or rules or regulations"? The WIPO Report, again, envisions that all claims by trademark holders that a domain name registrant registered an infringing name "in bad faith" be submitted to a single, uniform, worldwide dispute resolution process for adjudication.

Now, some, or even all, of these may be good ideas. But this is already way beyond the realm of technical "standards-setting," and we really must ask whether we really want or need this kind of global Internet policy and whether this is the way it should be put together. When did the affected constituency-all Internet users worldwide-decide that they want a global policy-making organ of this kind? Who decided that the bottom-up, decentralized, consensus-based governance structures under which the Internet grew and flourished are incompatible with its continued growth and development? When are we going to get a chance to ratify these new arrangements?

There are hard questions here, but one thing is clear; we need to disabuse ourselves of the notion that this is somehow not about Internet governance if we are going to make any serious headway on them. We know something about how institutions that possess life and death power can be constrained, about constitutions and constitutionalism, about the fragmentation of power and the need for checks on the exercise of power, and we better start thinking about this problem in these terms before too much more time elapses.

Oh and about James Madison? Madison not only thought more clearly and more insightfully about these questions than anyone before or since, he understood the necessity for public discussion and debate about issues of this kind; the Federalist Papers, in which he and Alexander Hamilton (and a somewhat recalcitrant John Jay) laid out the arguments for (and against) the constitutional structure put together in Philadelphia in 1787, began life, let us not forget, as newspaper columns appearing weekly in the New York press. We could do worse than to start thinking about updating that for the new cyber world we are building now.

Note: This document can be found on the web at the ICANNWatch website,

Draft Yokohama Statement on Civil Society and ICANN Elections - 27 June 2000

Internet Democracy Project

This draft document articulates a civil society perspective on the Internet Corporation for Assigned Names and Numbers (ICANN) and identifies issues for the upcoming At Large elections.

Civil society is a third sector of society alongside the state and the market. Civil society supports freedom of association, freedom of expression, participatory democracy, and respect for diversity. A vigorous civil society is also an important limit on the power of governments and on the power of the commercial sector.

We encourage individuals and organizations to discuss this statement internally and with others and to suggest improvements. This statement will be further developed at the:

Civil Society Forum
Yokohama, Japan (ICANN Meeting)
Thursday, 13 July 2000, 9:00-12:00
For more information see web sites above. Comments can also be submitted to Computer Professionals for Social Responsibility (CPSR) at <> .

Initial signatories (individuals):

Karl Auerbach - Individual Domain Name Holders Constituency (USA)
Chris Bailey - Internet Rights Campaign, Association for Progressive Communications (APC) (UK)
Tracy Cohen -LINK/Wits University (South Africa)
Marc Holitscher - Unit for Internet Studies (Switzerland)
Tomoya Inyaku - JCA-NET (Japan)
Hans Klein - Computer Professionals for Social Responsibility (CPSR) (USA)
Norbert Klein - Open Forum of Cambodia (Cambodia)
Veni Markovski - Internet Society - Bulgaria (Bulgaria), ICANN Membership Implementation Task Force Chair for East Europe
Milton Mueller - Syracuse University (USA)
Toshimaru Ogura - Net-workers against Surveillance Task-force (NaST) (Japan)
Nii Quaynor - Network Computer Systems (NCS) (Ghana)
Roberto Roggiero - INTERCOM - Ecuanex (Ecuador)
Marc Rotenberg - Electronic Privacy Information Center (EPIC) (USA)
Barry Steinhardt - American Civil Liberties Union (ACLU) (USA)
Shinji Yamane - Japan Chapter preparatory committee (Japan), Computer Professionals for Social Responsibility (CPSR)

Guiding Values

ICANN must be representative.

ICANN must be transparent.

ICANN must use bottom-up processes.

Intellectual property rights are not privileged over other rights.

ICANN must limit itself to technical policy-making.

The domain name space is not an exclusively public resource.

Artificial scarcity and centralization should be avoided.

ICANN must respect privacy.

Costs should be minimal and equitable.

Issues in the ICANN Elections

ICANN must be representative.

ICANN currently suffers from a democracy deficit. Since its creation in 1998 and continuing to the present, the commercial sector has had disproportionate representation on the Board of Directors. The democracy deficit will continue at least until all At Large Board seats are filled by elected representatives.

All At-Large Board seats should be filled by election as quickly as possible.

Any policies passed by a Board that is not fully representative should be subject to an annual vote of reauthorization ("sunset provision").

ICANN should embrace the membership provisions of its bylaws. Election procedures should be made more open, barriers to candidacy reduced, and full rights of membership should be recognized. In particular, provisions that attempt to weaken the legal rights of members should be removed from the by-laws.

The Domain Name Supporting Organization (DNSO) should restructure its constituencies to reduce the disproportionate representation given to business and intellectual property interests.

The DNSO should recognize new constituencies, including an Individual Domain Name Holders constituency, a developing countries constituency, and a small business constituency.

2. ICANN must be transparent.

Information-sharing should be maximized before, during, and after all ICANN decisions.

ICANN should make available records of the process of all decisions, except those pertaining to personnel or to the negotiation of contracts.

The cash flow structure of ICANN should be made public. A strict lineal path should be established between expense request, authorization, issuance of purchase order, receipt of invoice, delivery, and payment.

ICANN should publish a report with each of its decisions that explains how the action being taken fits within ICANN's scope and how the decision was created by an open and transparent process based on the consent of a majority of ICANN participants.

3. ICANN must use bottom-up processes.

ICANN is in danger of becoming an organization whose policies and practices are determined by its staff. ICANN needs to rededicate itself to its original conception as a decentralized, bottom-up standards making organization.

ICANN staff and CEO must show more respect for procedural safeguards and checks and balances.

The unelected Board seat reserved for ICANN's President should be eliminated. The entire Board should be democratically elected.

ICANN should not select a new President until after the first round of At Large elections. Any candidate for the position of ICANN President should not accept an offer until after the At Large elections.

No person or entity that played an active role in the creation of ICANN should obtain any benefit from ICANN or be a party to any contract with ICANN until 24 months have elapsed after that role has ceased (no "revolving door" of personnel transfers between ICANN and external partners.)

No person who has been member of the board or has held an executive office under ICANN should obtain any benefit from ICANN or be a party to any contract with ICANN until 24 months have elapsed after that role has ceased.

4. Intellectual property rights are not privileged over other rights.

The European Convention on Human Rights states, "Everyone has the right to freedom of expression" (Article 10). ICANN should not compromise the right of expression in order to protect the right of property.

DNS administration should not be leveraged to expand the scope of intellectual property rights (IPR). Civil law has been an adequate vehicle for regulating property. Changes in the scope and nature of international IPR protection should be made through national legislatures and international treaties.

Should DNS policy unavoidably intersect other policy areas, ICANN should be equally mindful of rights, laws, and norms protecting free expression, privacy, the public domain, and noncommercial use.

The Uniform Dispute Resolution Policy (UDRP) passed in 1999 without the representation or consent of Internet users should be subject to review and vote of reauthorization.

Intellectual property rights are best protected by establishing special zones in the domain name space for trademark (e.g. ".trademark").

5. ICANN must limit itself to technical policy-making.

The power over Internet users inherent in DNS administration should not be used to make public policy.

IP address management and DNS root server management need not be combined in the same organization. There are strong political, organizational, and technical reasons to separate address management from DNS policy making.

ICANN must not be used as an instrument to promote policies relating to conduct or content on the Internet. Its by-laws should explicitly recognize limitations on its powers in order to guard against expansion of mission ("mission creep").

6. The domain name space is not an exclusively public resource.

The assertion that "the [domain] name space is a public resource" (by ICANN's Governmental Advisory Committee) provides a basis for excessive state control. Likewise, the "natural monopoly" model of country code TLD (ccTLD) registries creates an opportunity for excessive control.

The domain name space is not an exclusively public resource. Assertions of public control over zones in the domain name space need explicit justification.

Public resources in the domain name space need not be under the control of national governments.

Multiple, parallel, and possibly overlapping TLDs registries for supra-national, national, sub-national, regional, cultural, linguistic, and other social and political groupings should not be excluded from the root. This is the basis of a vibrant civil society.

7. Artificial scarcity and centralization should be avoided.

Control points and artificial scarcity in DNS create barriers to Internet access and foster regulation of users.

The DNS root is a single point of failure on the Internet that threatens operational stability.

The single DNS root is a control point. ICANN should support the evolutionary development of the DNS away from a centralized architecture.

ICANN should encourage the interconnection of the DNS with alternate name spaces.

Scarcity in domain names creates opportunities for control. Expansion of the domain name space through the creation of new TLD registries should be ICANN's highest priority.

Expansion of the Internet domain name space should be unconstrained (except for technical constraints -- to the extent that such constraints exist.) Expansion through decentralization of the root and growth in top level domains is especially desirable.

The use of domain names as a marketing device to index content creates excessive value in domain names and creates disincentives to innovation. The technical evolution of DNS should not be unduly inhibited by its use as a marketing technology by commercial users.

8. ICANN must respect privacy.

ICANN's policies and internal procedures should adhere to Fair Information Practices, based on the OECD Privacy Guidelines.

ICANN's policies for domain name and address management should not discourage the adoption of genuine privacy enhancing techniques or undermine the right of anonymity.

9. Costs should be minimal and equitable.

Similar services delivered in different parts of the world can have different cash values. Likewise, users' ability to pay can vary dramatically.

ICANN's costs should be distributed in a manner that corresponds to the costs caused by different users.

Many costs have arisen from the high priority given by ICANN's Board to address the concerns of commercial Internet users. Costs assessed to those users should reflect this.

ICANN should at all times strive to minimize costs (e.g. rather than holding Board meetings in business class facilities, ICANN should use non-profit quality facilities.)

ICANN should allow an outside audit of expenses, business practices, cost controls, and accounting methods. The standard of evaluation should be that of public-benefit non-profit entities rather than those of for-profit corporations.

Business' case against a union TLD

July 8, 2000

Internet Corporation for Assigned Names and Numbers (ICANN)
4676 Admiralty Way, Suite 330
Marina Del Rey, CA 90292
Re July 2000 ICANN Yokohama Meeting Topic Introduction of New Top-Level Domains

To Whom It May Concern

LPA is pleased to submit comments regarding the likely addition of new top level domain names by the Internet Corporation on Assigned Names and Numbers (ICANN). In particular, LPA registers its strong opposition to the adoption of a .union chartered top-level domain (TLD), which was mentioned as an example of a new domain name in ICANN's June 13, 2000 background document. The addition of a .union TLD would cause undue confusion among Internet users, particularly among employees who are not represented by a union. Moreover, a .union TLD would violate many of the principles announced by Working Group C in its supplemental white paper. At a minimum, LPA recommends that ICANN refrain from accepting a .union TLD in the initial round of expansion and instead draw on the lessons learned from the implementation of other chartered top level domains that are added before evaluating the viability of .union TLD.

LPA is an association of the senior human resource executives of more than 200 leading corporations in the United States. LPA's purpose is to ensure that U.S. employment policy supports the competitive goals of its member companies and their employees. LPA member companies employ more than 12 million employees, or 12 percent of the private sector U.S. workforce. LPA members have a substantial interest in making sure that employees receive accurate information about whether they are represented by a union or not. The addition of a .union top-level domain could prematurely undermine employee confidence in the use of the Internet as a tool for communicating with employees.

I. The Creation of a .union TLD Would Create Confusion and Administrative Problems

LPA believes that a .union TLD would create more problems than benefits and should be dropped from consideration, particularly at this stage of domain name expansion. The creation of a .union TLD could confuse employees, especially those who are not familiar with union organizing procedures. The .union domain name would also likely require the chartering entity to put in place sophisticated application and management procedures to reduce the inevitable disputes that would arise among competing unions. Moreover, it is far from clear that a .union TLD would pass muster under the principles set forth by the Domain Name Supporting Organization's Working Group C.

A. Background on the .union Proposal

The background paper posted on ICANN's web site on June 13, 2000, used a .union top level domain name (TLD) as an example of a non-commercial chartered TLD that could be sponsored and managed by a group of interested labor unions.1 The .union TLD proposal was introduced by Jamie Love, the director of the Consumer Project on Technology and John Richard, director of Essential Information, both technology-related interest groups affiliated with Ralph Nader.

On March 1, 2000, they sent a letter to Esther Dyson, the chair of ICANN, indicating their interest in establishing a .union TLD, described as "a 'union label' for cyberspace. Use of the domain would be restricted to bona fide labor unions. Examples of the use of this TLD would include nike.union, exxon.union, microsoft.union, as well as other uses . It is our goal to use the .union domain to strengthen union organizing efforts, and to make it easier for workers at a firm to communicate with unions that represent workers at the firm, or who are seeking to organize workers at the firm, and for unions in different countries to coordinate efforts with each other." (Letter from James Love and John Richard to Esther Dyson, chair of ICANN, Mar. 1, 2000.)

Shortly after the letter was sent, a meeting was held reportedly involving Mr. Love, union leaders in the United States, the chair of ICANN's working group on trademark issues and senior representatives from the Department of Commerce to discuss the prospects of a .union TLD. During the meeting, some union representatives noted that having domains under a company name could enhance cooperation among unions in several countries in dealing with the same company. Nothing in the written account of the meeting mentioned the prospect of employee confusion.(2)

Subsequent to the meeting, a posting from Pruett Duncan of the International Confederation of Free Trade Unions indicated that many of the international trade union organizations would be interested in participating in, and even managing, a .union TLD. (3) This indicates that already there is competition for the management of a .union TLD between the AFL-CIO in the United States and a group of international trade union organizations. Competition for sponsorship of a .union TLD aside, LPA believes that a .union TLD would create excessive confusion among employees and would be counterproductive in enhancing the Internet in the initial round of domain name expansion.

B. The .union Proposal Would Create Employee Confusion

LPA's primary concern with a .union TLD is that a domain name such as nike.union would create substantial employee confusion and labor relations problems. These issues take a .union TLD proposal out of the "technical" realm occupied by ICANN and into the realm of labor relations policies governed by legislative and administrative bodies in the United States and other countries. A .union TLD is also likely to create several difficult domain name management issues that ICANN would have to resolve with the organization that is granted a .union TLD charter.

A company.union TLD could potentially confuse employees into thinking that they were already represented by a union that was formed by or recognized by the company. In the United States and other countries, it is illegal to have a union that is formed by the company (a company union).4 Generally, in the United States, an employer can only recognize a union after employees have voted for a union in a secret ballot election or after the company has obtained an objective measure that a majority of employees in the bargaining unit support the union. The union normally demonstrates a showing of interest by having employees sign union authorization cards that indicate that the employee wants to be represented by a labor union.

Even though a company union is illegal in the United States, a nike.union domain name, for example, could create the impression that all or many Nike employees are already represented by a union initiated or run by the company (i.e., the Nike union). Alternatively, a nike.union domain name could lead employees to the conclusion that they already had an independent union, and could allow a union organizer to play upon this impression to convince employees to sign union authorization cards. There is substantial evidence in U.S. labor jurisprudence that individual union locals seeking to organize particular employers have misrepresented the significance of signing an authorization card in order to convince employees to sign the cards. (5) If a .union TLD were created, at the very least, the entity holding the charter would have to set well-defined rules regarding the accuracy of the information available on chartered web addresses.

Further confusion could arise for employees working at different companies that have the same name. As John Berryhill pointed out in a March 25, 2000 posting, "at delta.union will they be the faucet makers or the airline employees?" (6) One can think of several other examples that could further lead to employee and union confusion.

C. A .union TLD Would Require Constant Oversight by the Chartering Entity

A chartered .union TLD would require significant monitoring and oversight by the holder of the charter, and it would require a sophisticated dispute resolution mechanism to resolve conflicts that develop among unions, regarding rights to domain names and misinformation posted on web sites. A .union TLD would be used primarily for adversarial and advocacy purposes, namely to convince employees to join one of a number of competing unions or to engender support among employees for a union. Thus, although a .union TLD may be non-commercial, it is not non-commercial in the same sense as a .museum TLD or a .edu TLD, which are primarily informational in nature. The ICANN background document erroneously grouped all three TLDs together.

A .union TLD also could create significant conflict between and among unions, depending upon how the charter is managed. For example, assuming unions are the only entities eligible to register second level domain names under a .union TLD, ICANN would need to determine

- whether an individual union would have the right to register a company.union domain name or whether all unions interested in representing employees at the company would have a right to post information on the site;

- if an individual union was allowed to register a company.union domain name, how the union would demonstrate that it had a legitimate interest in representing employees at the employer;

- if an individual union was allowed to register a .union domain name, the point in the organizing process at which the union would be allowed to register the name (before organizing began in earnest, after the union has begun organizing, after a showing of interest by employees);

- if an individual union was allowed to register a .union domain name, how the chartering entity would prevent "cybersquatting" among unions.

If ICANN allowed the chartering entity to sell a company.union domain name to an individual union, other unions would be excluded from posting information on that web site. Significant disputes among unions competing to represent employees at the company would result. For example, what would happen if a competing union secured the rights to company.union, where company's employees were already represented by a different union? Would the chartering entity be required to grant a right of first refusal to the union that represented employees at the company? In addition, where multiple unions represent different groups of employees at a company, such an arrangement would exclude unions that have legitimate interests at the company.

Matters would be complicated further under such an arrangement if employees voted to decertify a union as the exclusive bargaining representative of the employees at the company. After the decertification, would the chartering entity require the owner of the do main name to surrender the name? What if the employees decertified one union and elected to be represented by another?

Would the decertified union be required to transfer ownership over the domain name to the newly certified union? These and other complex issues would be created if individual unions were allowed to own a company.union domain name.

Based on the above concerns, it would appear that, in the interest of fairness, a .union TLD charter would have to allow all interested unions to post information on a company.union site to be equitable. Yet, this "gateway" approach presents its own problems. The chartering entity would still need to determine when a union would be allowed to post information on a company site. In larger companies that have many unions, this could become excessively complicated. It could be possible to have dozens of union locals posting information to a single company site, creating more confusion for employees and imposing significant administrative burdens for the chartering entity, which arguably would be required to monitor the postings.

In addition, the chartering entity would presumably be required to establish dispute resolution procedures to resolve claims made by competing unions about inaccurate information posted on the web site by their rivals. Such procedures could lead to a significant new bureaucracy to monitor and resolve claims as unions compete for new union members.

D. The Proposal Would Violate the Principles for New TLDs Set Forth by Working Group C

The Domain Name Supporting Organization's Working Group C on new TLDs approved five additional principles for assessing new general TLDs which are pertinent to the discussion of a .union TLD. These are meaning, enforcement, differentiation, diversity, and honesty. As analyzed below, whether a .union TLD would meet several of these principles is questionable at best and points to the conclusion that a .union TLD should be shelved.

** Meaning. According to the Working Group C summary, "an application for a TLD should explain the significance of the proposed TLD string, and how the applicant contemplates that the new TLD will be perceived by the relevant population of net users. The application may contemplate that the proposed TLD string will have its primary semantic meaning in a language other than English." (7)

As explained above, the relevant population of net users for a .union TLD would be employees, who could be confused by a company.union domain name. By having a company.union TLD, employees may be led erroneously to believe that they were already represented by a union that was recognized by a company.

** Enforcement. "An application for a TLD should explain the mechanism for charter enforcement where relevant and desired." (8) Enforcement of the .union TLD charter has several aspects to it. Enforcement of a charter typically means that the charter is available only to those entities that are interested in the subject matter in question, in this case, labor unions. Enforcement also means, however, the need to ensure that all unions that wish to provide information under a .union domain name may do so. Because many unions can compete to represent the same employees at a company or can represent different employees at the same company, enforcement is likely to be a substantial undertaking that would require significant oversight by the chartering entity.

** Differentiation. "The selection of a TLD string should not confuse net users, and so TLDs should be clearly differentiated by the string and/or by the marketing and functionality associated with the string." (9) As noted above, the use of a company.union TLD could confuse some employees into thinking that their employer had recognized a union when it had not or that the employer had recognized a union other than the one that represented its employees.

** Diversity. "New TLDs are important to meet the needs of an expanding Internet community. They should serve both commercial and non-commercial goals." (10) Although a .union TLD would serve non-commercial goals, the ICANN background information significantly misstated the effect of union websites by characterizing them as similar to .edu or .museum. Union web sites are information and advocacy tools in union-management relations that are inherently adversarial. LPA believes that there is substantial opportunity for some unions to profit at the expense of others. LPA suggests that diversity and competition would be better protected by requiring unions to register through existing TLDs or new general and open TLDs. In addition, many unions are now providing benefits that involve commercial transactions. For example, many unions offer discounts on credit cards, computers and Internet access. These transactions would clearly take the web sites out of the non-commercial realm.

** Honesty. "A TLD should not unnecessarily increase opportunities for malicious or criminal elements who wish to defraud net users." Even if intended for bona fide information dissemination and sharing, a .union TLD creates increased opportunity for unions to convince nonunion employees that they are already represented by a union by virtue of the web site name. As stated above, some unions could use the existence of a company.union domain name to convince employees that the company supports a union and to convince employees to sign union authorization cards to demonstrate their support for the union.

In sum, LPA believes that the principles suggested by Working Group C support the notion that a .union TLD could be counterproductive to the domain name system.

III. More Time Is Needed to Consider New Chartered, Non-Commercial TLD Proposals

LPA does not involve itself with the technical aspects of web management, but it does believe that ICANN should provide additional time for public comment on proposals for new chartered, noncommercial TLDs. Specifically, LPA believes that ICANN should provide a minimum 60-day period for evaluating proposed new chartered, non-commercial TLDs after the deadline for submission has passed, instead of the one-week period currently suggested. This puts all parties interested in new chartered, non-commercial TLDs on a level playing field, rather than disadvantaging those who are interested in proposals that were posted at the last minute.

LPA also recommends that applicants be required to provide a detailed account of how the application process for new domain names would operate under chartered domains. Applicants should be required to demonstrate how they will resolve competing claims where more than one party is likely to want access under each domain name.

IV. Conclusion

LPA believes that the creation of .union TLD is unnecessary and will create more confusion and conflict than access to useful information. A .union TLD is likely to cause infighting among unions, require the chartering entity to keep a vigilant watch on how domain names are used, and require regular dispute resolution proceedings. Above all, the prospect for employee confusion merits that the proposal be dropped at this stage of Internet expansion. Finally, to ensure a full airing of the views of interested parties, LPA recommends that ICANN adopt a 60-day evaluation and comment period following the deadline for filing proposals.

Sincerely yours,

Daniel V. Yager
Senior Vice President and General Counsel <http//>
SUITE 1200
WASHINGTON, DC 20005-2605
TEL 202.789.8670
FAX 202.789.0064

1 See ICANN Yokohama Meeting Topic Introduction of New Top-Level Domains, II(C)(2-3) <<http//> topic.htm#IIC2>.

2 James Love, Friday's Dot Union Briefing, April 1, 2000, <<http//>.

3 Duncan Pruett, Some Thoughts on a Chartered GTLD for Trade Unions, May 31, 2000, <<http//>.

4 See 29 U.S.C. 158(a)(2).

5 Cases in which unions have used misrepresentations to convince employees to sign authorization cards include Nissan Research & Dev. Inc., 296 N.L.R.B 598 (1989) Burlington Indus., Inc. v. NLRB, 680 F.2d 974 (4 th Cir. 1982); Medline Indus., Inc. v. NLRB, 593 F.2d 788 (7 th Cir. 1979); Fort Smith Outerwear, Inc. v. NLRB, 499 F.2d 233 (8 th Cir. 1974); Southern Cal. Associated Newspapers, Inc. v. NLRB, 415 F.2d 360 (9 th Cir. 1969); Schwarzenbach-Huber Co. v. NLRB, 408 F.3d 236 (2d Cir. 1969); J.M. Machinery Corp. v. NLRB, 70 L.R.R. M. 3355 (5 th Cir. 1969); Lenz Co. v. NLRB, 396 F.2d 905 (6 th Cir. 1968); Southland Paint Co. v. NLRB, 394 F.2d 717 (5 th Cir. 1968); Dan Howard Mfg. Co. v. NLRB, 390 F.2d 304 (7 th Cir. 1969); Swan Super Cleaners, Inc. v. NLRB, 384 F.2d 609 (6 th Cir. 1967); Dayco Corp. v. NLRB, 382 F.2d 577 (6 th Cir. 1967); Nichols-Dover, Inc. v. NLRB, 380 F.2d 438 (2d Cir. 1967); Eng'rs & Fabricators, Inc. v. NLRB, 376 F.2d 482 (5 th Cir. 1967); Freeport Marble & Tile Co. v. NLRB, 367 F.2d 371 (1 st Cir. 1966); Bauer Welding & Metal Fabricators, Inc. v. NLRB, 358 F.2d 766 (8 th Cir. 1966).

6 John Berryhill's Response to James Love Regarding a .union Top-level Domain Name, Mar 25, 2000, <<http//>.

7 Addendum to WG-C consensus report to the Names Council, Apr. 17, 2000, <http//> addendum.html.

8 Id.

9 Id.

10 Id.

About the author

Eric Lee, the Information and Communications Technology Co-ordinator of Labour and Society International, founded the popular LabourStart website. He is the author of three previous books, including The Labour Movement and the Internet: The New Internationalism (Pluto Press, 1997). He lives in London, is married, and has two sons.