This study considers to what extent the Internet Corporation for Assigned Names and Numbers (ICANN) has achieved its stated goal of a "representative" and "open" decision-making process. An initial review of approximately 100,000 postings by members of the Internet user community to ICANN's e-mail lists and public online forums — indeed just a few of the several means of public participation in ICANN's decision-making process — showed that public commentary for or against a given proposal before the Board does not correlate strongly to an outcome either for or against that proposal. The data suggest that the Board has been more likely to rely heavily upon staff recommendations and upon the input of the Supporting Organizations, in which the public can also participate, than on broad-based input from the Internet user community.
In large measure because of the difficulty of assessing public involvement in the Supporting Organizations, it is unclear whether ICANN has been successful at collecting commentary that is representative of the Internet user community through its full range of user input methods. On some key issues, such as the reassignment of the .org domain and on specific Top-Level Domain extensions, Internet user community members posted very few substantive comments, despite the very large aggregate number of postings, to the public message boards.
In a parallel analysis of the Board votes at ICANN, we found that, of the 723 resolutions passed by ICANN, 597 were passed without a dissenting vote. Only 10 of the 42 directors in ICANN's history cast a single "nay" vote during their tenure on the Board.
None of the findings in this report bear on whether or not the ICANN Board made the right decision on any given issue from the perspective of fulfilling the technical coordination — rather than the procedural — side of its mission.
Based upon these findings and related evidence, we reach three conclusions:
1. ICANN's experimentation in new modes of corporate governance has broadly failed in terms of attracting and incorporating "representative" input from the global Internet user community, at least with respect to the public online forums. This failure underscores the need for an overhaul of ICANN's governance structure away from its semidemocratic past. Online message boards and public e-mail lists have proven extremely limited for ICANN as a means of engaging the global Internet user community in the decision-making process.
2. Regardless of the new structure adopted, ICANN should clarify the way(s) in which users can involve themselves in the decision-making process for managing the domain name system, at a minimum by indicating plainly the relative weight given to Supporting Organization input as compared to other forms of direct public input.
3. To the extent that we seek new means of governing the technical architecture of the Internet, we ought to look beyond ICANN, which may never have been the right place for such experimentation given its limited technical mandate.
ICANN has repeatedly affirmed its commitment to carrying out a decision-making process that is representative, open and bottom-up. Imagine you were an individual Internet user interested in participating in ICANN. You go to ICANN's publicly accessible web site and find the page entitled "Participate in ICANN". ICANN's first suggestion to you, near the top of the page, is to participate in one of ICANN's many public forums, and you are directed to forum.icann.org. (ICANN also suggests on this page that you participate through one of the Supporting Organizations or the At Large Advisory Committee, neither of which we have studied closely in preparing this report.) This study seeks to quantify and contextualize the extensive data available at forum.icann.org to determine whether or not public participation through these online forums has been taken into account by the ICANN Board in its decision-making.
Since 1998, the Internet Corporation for Assigned Names and Numbers (ICANN) has been responsible for the technical coordination of the Internet's Domain Name System (DNS). The DNS enables people to locate resources on the Internet with textual identifiers like www.harvard.edu in lieu of the more difficult to remember IP addresses that identify each computer on the Internet. The effects of ICANN's actions can be found in the low registration fees for .com, .org, and .net domain names and in the availability of both new general top level domains (gTLDs) like .info and .name and country-code top level domains (ccTLDs) for use by citizens and organizations based in countries around the world. Recently, ICANN facilitated VeriSign's implementation of a controversial waiting-list service for registered domain names. ICANN also played a major role in the effort to resist VeriSign's initiative to redirect all traffic intended for nonexistent .com domains to a search site run by the redirecting registry. While most Internet users are not familiar with the institution, ICANN undeniably plays an important role in maintaining the reliability and stability of the Internet.
ICANN's organizational structure is a complex, experimental form. This complexity is a high initial hurdle to those who wish to become involved in ICANN's work, as someone approaching this decision-making process might have a difficult time understanding how to participate meaningfully and what forms of input carry weight with the Board. In a parallel working paper, co-author John Palfrey explores the shortcomings of ICANN's "semidemocratic" organizational structure, as well as ICANN's global Election of 2000, in connection with the data presented in this report.
In the first section of this paper, we discuss the diversity of channels available to those members of the Internet user community interested in participating in ICANN and explain why we chose the ICANN mailing list archives and the bulletin boards at forum.icann.org bulletin boards as our primary research environment.
In the second section of the paper, we introduce the methodology used to analyze the Internet user community's opinion as it is expressed on the mailing lists and forum.icann.org.
In the third section of this paper, we look at the shape of the data collected from both the mailing list archives as well as forum.icann.org and ask whether these channels have the potential for successful public discussion.
In the fourth section of the paper, we analyze the public's opinion on four issues representing four types of decisions that ICANN makes. Along with the new quantitative data, we also include a discussion of how we perceived the ICANN board incorporating those opinions into its decision making process for each issue.
Finally, in the fifth section of the paper we discuss the implications of our results and suggest new avenues of research that would build upon these preliminary results.
ICANN purposefully created many channels through which the public could observe and participate in its decision-making process.
ICANN held an online global election in the Fall of 2000 . This election was the most ambitious of its several attempts to engage the world's Internet users in a common, open, democratic process . Despite severely limited resources, ICANN ran an online election that enfranchised voters of every region of the world and drew more than 75,000 people to register to vote at the Internet-based polling booths set up by election.com for the purpose of electing five ICANN directors . The relative success of this election at a tactical level is far overshadowed by the failure of ICANN's experiment in At Large representation and participation. The story of formal representation of the public on ICANN's board effectively drew to a close on December 15, 2002 as the elected directors' terms ended and the by-laws changed to end the experiment . The representatives elected to the ICANN board in 2000 had quite a different voting pattern than Board members appointed through other means (please see our analysis of this issue).
Since 1999, ICANN has held between three and four public meetings each year in locations around the world. Those who are able to appear in person at the public meetings are allowed to attend and participate without charge. Those unable to attend these large meetings in person are invited to participate remotely, also free of charge, where technologically and procedurally feasible. Public participants are not, for instance, inhibited from wandering in and out of most committee meetings, but ICANN has not opted to broadcast each of these small-group sessions over the Internet. During large public forums, remote audiences have historically been able to watch the proceedings via webcast and to participate in these forums using a variety of software solutions. In the past, individuals were able to chat about the proceedings via IRC, email comments and questions to a staff person present at the meeting who would present certain relevant questions to the Board's Chairman, watch real-time and archived feeds of the events, and review transcripts and "scribe's notes," among other descriptions of the proceedings. Over the years, the precise methods available for public interaction have changed and recently been reduced — IRC, for instance, has not been seen at recent meetings, nor have the near-instanteous posting of meeting notes; full archival presentations of proceedings, including questions received through real-time channels but not read to the in-person group; and prompt, indexed posting of video and audio footage of meetings. ICANN's commitment to the openness of access to most public meetings has remained substantial, though not at a constant level .
ICANN has traditionally had a series of influential Supporting Organizations that "assist, review and develop recommendations on Internet policy and structure within three specialized areas" . Each of the three current Supporting Organizations names three members of the ICANN Board. Each Supporting Organization welcomes the involvement of the public. Participation in a Supporting Organization is likely the most effective means of becoming involved in the ICANN decision-making process. ICANN has also introduced the At-Large Advisory Committee ("ALAC") and the Nominating Committee ("NomComm"), in which members of the Internet user community may participate. (The most glaring shortcoming of this report is that we have considered, but not found a reliable way fully to assess, public participation through the Supporting Organizations and committees such as the ALAC and the NomComm.)
Many of the discussions that lead to substantive policy-making at ICANN take place in small meetings and teleconferences of support organization committees or the ICANN Board of Directors. Global teleconferences take place with some frequency. While most of these teleconferences are not open to the public, transcripts of these teleconferences, and related small face-to-face meetings are often released shortly after any substantive discussions take place. At these small gatherings, public participation is minimal in real-time, but ICANN has nonetheless expended substantial effort to ensure that the decisions are made in a relatively transparent environment. For example, the minutes of all board meetings are available on the Internet for public review.
ICANN has continuously offered several Internet-based channels for public participation. Our study focuses on two such channels: mailing lists and, to greater extent, the discussion boards on forum.icann.org. Our selection of mailing lists and the bulletin boards at http://forum.icann.org does not give us a complete view of the Internet user community's opinion on any given matter, but these data are readily accessible to any observer and originate in a space broadly recognized as an appropriate place to comment on ICANN's policy proposals.
Official ICANN mailing lists include both closed lists, such as the Name Council List, which has a restricted membership, and open lists, such as the GA-Full list, in which anyone with an email address can participate. The diversity of email formats prohibited us from easily using the mailing list data to analyze opinions on particular issues. Nonetheless, we downloaded and sorted all available archives mailing list messages in an attempt to gain an empirically based understanding of the suitability of the space as a channel for public participation.
Forum.icann.org, like the open mailing lists, creates a space on the Internet where the public can comment on current ICANN issues. We highlight two differences between forum.icann.org and the mailing lists. First, forum.icann.org has no formal membership and is a passive space. Unlike a mailing list, one cannot "subscribe" to forum.icann.org. The only way to monitor the space is to spend time watching and reading the web page. Second, forum.icann.org is organized by topics. When ICANN produces a document, makes a proposal, or takes an action, it is customary to create a corresponding page on forum.icann.org where the public can comment on the move. This creation of issue specific forum pages provides easy access to a self-sorted, issue-specific database of public commentary of certain matters before ICANN. The self-selection process leading to forum participation skews the expressed opinions, but ICANN's actions are only a topic of interest to a limited population. Therefore, we decided to analyze these data with the caveat that trends found on forum.icann.org likely do not correspond to trends found outside the forum.icann.org pages. (This caveat points, once again, to the primary limitation of this report, which is its limited scope and its lack of inclusion of meaningful public participation data from within the Supporting Organizations and other ICANN committees.)
Following the selection of publicly archived mailing lists and forum.icann.org as the focus of our research, we wrote a collection of perl scripts that systematically downloaded and sorted approximately 100,000 mailing list emails and forum comments. The data we gathered do not reflect all discourse that took place in the mailing list and forum.icann.org spaces; some archives are no longer available online and human error may have led to the omission of other undiscovered data sets. We collected data from all of the forum.icann.org web-pages published before 1 July 2003 and all of the mailing lists with archives on the DNSO/GNSO web pages. One researcher traveled to Montreal, Canada, to participate in the public meeting that took place during the course of this research project.
After downloading and sorting the data, we graphically inspected the data from a variety of sub-populations to find out whether a random sample was appropriate. We have included a short discussion of these findings in the Channel Analysis section of this paper.
In analyzing each of our four target issues, we decided at the outset whether it was feasible to analyze every available comment within our population of interest or to sample from the population. When fewer than 400 comments were made regarding an issue, we included all of the comments in our analysis. When more than 400 comments were posted, we randomly sampled from the population of comments. Sample sizes were calculated at 95% confidence levels with a confidence interval of plus/minus 5%. If a distribution was particularly skewed — for example, when a few users post a large portion of the comments — we sampled the data twice: once for the opinions of the majority of users and once for the opinions of the heavy users. The sample sizes for these bifurcated populations were calculated at the same power levels as the original sample sizes. For all of the instances where this was a concern, we discovered that the users dominating the conversation were predominantly posting off-topic messages or "flames" irrelevant to the serious consideration of a given matter. We have sought to include only comments that we deemed to be substantive in the presentation of our issue analysis data.
Researchers familiar with the issues, but generally unfamiliar with the individuals posting comments to the message boards, coded each comment within the context of the issue being analyzed. What the rater looked for in each post directly reflected the issue that was being discussed on that particular board. If, for example, the message board was intended to gauge support for the implementation of VeriSign's proposed Wait-Listing System, comments were tabulated as supportive of the proposal, in opposition to the proposal, or off-topic. If a rater experienced difficulty determining the intention of the posting user, a second rater was brought in to collaborate and, in all cases, consensus was reached as to the intention of the user. We performed a quality-assurance check of our data analysis, in which a researcher other than the initial researcher reviewed the raw and compiled data. Prior to this publication, John Palfrey presented draft, initial findings of these data in three settings: before a group of Internet law professors in August, 2003; at a seminar on Internet governance at the Kennedy School of Government in October, 2003; and at a research meeting of Berkman Center for Internet and Society at Harvard Law School's fellows in November, 2003.
In addition to our quantitative analysis of the opinions expressed in forum.icann.org for each selected issue, we briefly summarize the issue history and explain how ICANN's Board may have incorporated the public commentary, along with other inputs such as recommendations from the Supporting Organizations (SOs) and others, into its decision-making process.
Whether a low-traffic closed mailing-list or heavy-traffic open forum, a similar distribution was found when the number of comments made (x) were graphed against the number of users posting that number of comments (y) to a particular forum or mailing list.
When the number of posts (x) was graphed against the number of users posting that number of posts (y), graphs similar to this one were often found, regardless of the population analyzed.
The data show a Pareto distribution. Participation in these forums is characterized by a large number of postings by a few persons and a small number of postings by a large number of persons. These highly skewed graphs straighten out when graphed on a log-log transform scale.
The distribution of these data provides some insight into the viability of these channels as spaces for public discourse. In all of the spaces we examined that existed long enough for a community to form (certain sub-populations of forum.icann.org were too small or time-limited), a small subset of the users dominated the discourse. On forum.icann.org, for example, one-percent of the users contributed 35-percent of the content on the bulletin boards. Other populations showed even more lopsided distributions. By completely dominating the discussions, the most vocal participants minimize the visibility of other user's opinions and maximize the visibility of their own agenda. However, recipients might begin to screen out the frequent poster's commentary once the substantive nature of those comments becomes repetitive or stop monitoring the discourse entirely.
For different categories of data, we identified who the most vocal participants in each space were. While the most frequent posters in each list contributed a significant proportion of the bandwidth, the heavy users were not all alike. A certain type of heavy user dominated each of the list categories. We identified three types of discussion space for our analysis: (1) Closed organization-oriented lists, (2) Open debate lists, and (3) Open bulletin boards.
Our findings are shown in the tables below.
ICANN created dozens of message boards on forum.icann.org on various topics. We identified four types of board decisions and selected four issues meant to be exemplary of those respective types of decisions, set forth in the table below.
First, we considered the process for introduction of new top-level domains in 2000 . ICANN's initial introduction of new top-level domains was eagerly awaited and hotly contested, as 47 applicants submitted proposals for new TLDs . The proposed new TLD was evaluated via a "threshold" review to sort out those proposals that met initial technical and financial requirements. One of the criteria on which decision-making was to be based was public comment, which prompted this review. ICANN's Board considered each proposal against a series of criteria, ultimately choosing seven new TLDs. In some cases, new TLDs chosen had substantially more favorable than negative public commentary, as in the case of the .coop TLD. In others, such as the contested .biz consideration, ICANN's Board chose the only proposal, that of JV Team LLC (now NeuLevel), with more negative than positive public commentary. Certain proposals that met the threshold reviews and received more positive than negative commentary from the public, such as .dir and .geo, were not chosen as new TLDs. Finally, certain proposals with substantially more positive public commentary than negative commentary but which did not meet the threshold standards, such as .law, were not chosen as new TLDs. No particular pattern of support or non-support based on public commentary emerges from these data. Given the paucity of public comment related to most proposals, such an absence of a pattern is hardly surprising.
In November 2000, the ICANN Board of Directors selected seven new top-level domains (TLDs) to be introduced into the DNS. ICANN began soliciting applications for new TLDs in August 2000 and received 47 applications for consideration by 20 October 2000. Applicants paid a non-refundable fee of $50,000 to submit a proposal. The ICANN staff, with the assistance of outside experts, evaluated each of the applications with respect to a set of technical, functional, and managerial criteria, as well as more holistic factors like functional and geographic diversity. This round of applications was intended to yield only a small number of new TLDs in the interest of maintaining the overall stability of the DNS and evaluating the viability of new TLDs.
All submitted applications were first evaluated on the basis of a "threshold" review that determined whether each individual application met basic technical and financial requirements. Applications that passed this threshold review were then evaluated according to a set of predetermined criteria, with consideration given to public comments and category-specific factors.
Of the proposed TLDs, some received applications from multiple parties (eg. ".biz") while others werluated with respect to the managerial criteria.
Several of the uncontested applications received significant public support, such as .law (158 supporting, 1 opposed), .event (2 opposing, 20 supporting), and .mall (5 opposing, 12 supporting). Each of these applications, however, failed the threshold review.
Two applications (.dir [3 opposing, 9 supporting] and .geo [6 opposing, 16 supporting]) offered new services, met the threshold requirements, and received positive public support overall. Neither of these services was selected for further negotiations by ICANN.
One application, Association Monegasque des Banques (.fin), met the threshold review requirements but received significant negative response (5 opposing, 1 supporting) from the public comments. The ICANN Report cited these comments in determining that the applicant was not sufficiently representative of its target industry, in this case global banking and finance.
Both the .kids and .biz domains received multiple applications for consideration. As with the uncontested applications, a number of these applications were eliminated on the basis of a threshold technical and business review. From the .biz applications, Affinity Internet and Abacus America, Inc. were eliminated for failing to meet threshold standards. From the .kids applications, DotKids, Inc. was eliminated. All three of these applications received at least some public support (Affinity Internet: 5 opposing, 9 supporting; Abacus America, Inc: 7 opposing, 8 supporting; DotKids Inc.: 4 opposing, 15 supporting). In each instance, however, as in many cases related to the new TLD applications, the number of total comments was very small.
Of the remaining three applications for the .biz domain, none appeared to receive overwhelmingly positive or negative evaluations with respect to the August 15th Criteria examined in the ICANN Report. They did, however, receive markedly different treatment in the Public Comment Forum. The application from iDomains (3 opposing, 5 supporting) received more positive feedback than negative, while the applications for KDD Internet (4 opposing, 2 supporting) and JV Team (4 opposing, 1 supporting) received more negative feedback than positive. The total number of comments on these applications was extremely small, however, in absolute terms. Ultimately, the JV Team application was selected for further negotiations, despite the mostly negative commentary from the Public Comment Forum.
The remaining applications also received no overwhelmingly positive or negative commentary. Certain legal and administrative aspects of such a restricted content domain may also have led the ICANN staff to recommend that none of the applications be reviewed further.
Input from public comments does not appear to have played a major role in the final TLD selections made by the ICANN Board. Public commentary appears to have been considered as a secondary factor, perhaps reasonably so given the paucity of comments with respect to most applications. The effect of public comments at most appears to have been to enable certain applications to merit further consideration when they would otherwise have been eliminated outright (see e.g., Image Online Design's application for .info), and supporting decisions that might have been made regardless of such public input (see e.g., Association Monegasque des Banques). The threshold review process appeared to play a much more significant role in eliminating applications, and considerations other than public commentary appeared to play a significant role in final selections, given that positive commentary did not correlate strongly with selected applications.
The small number of public comments received related to many of the new TLD applications, and on other questions of importance before the ICANN Board, could lead to a variety of conclusions. First, the small number of posts could justify an argument by ICANN staff and Board members that the data are too sparse to be relied upon. Similarly, one might conclude that these public comments should be given little or no weight, given the ease with which an organizer might encourage a small number of persons to post commentary for or against a given proposal and skew the results. From the opposite perspective, one might conclude that the small number of comments means that ICANN has failed in soliciting meaningful input from the Internet user community at large — holding aside the question of whether the Supporting Organizations or other means of input could have provided such input — and as such has not lived up to its promise of a representative decision-making process. A final possible conclusion is that these public online comment forums and mailing lists are the wrong technologies for soliciting meaningful Internet user community involvement.
Second, we analyzed the process by which ICANN considered VeriSign's proposal for a new wait-listing service. During a special meeting of the Board on August 23, 2002, ICANN voted to implement the VeriSign Global Registry Services (VGRS) proposal. The public commentary was overwhelmingly opposed to the introduction of the wait-listing service. The Names Council, adopting a report developed by the DNSO Transfers Task Force, also recommended against this proposal. ICANN's general counsel mentioned the possibility of endangerment of legitimate business interests in the event of adoption of the wait-listing service proposal, but did not directly oppose the proposal. Despite the strong objections expressed by those who submitted public commentary and the opposition of the Names Council and the DNSO Transfers Task Force, the ICANN Board decided to support the VGRS waiting-list service proposal.
During a Special Meeting of the Board on 23 August 2002, the ICANN Board of Directors voted to implement the VeriSign Global Registry Services (VGRS) proposal to amend the registry agreement for the .net and .com TLDs for the purpose of conducting a yearlong trial of a wait-listing service (WLS). A WLS allows an individual to put her name on a reservation 'waiting list' for domain names currently registered through an ICANN-accredited registrar. When a domain name is deleted through the normal course of operations and all grace periods have elapsed, that domain name will be automatically subscribed to the individual next in the queue on the waiting list for that particular domain name. VGRS originally proposed a WLS amendment to the DNSO Registrars Constituency on 30 December 2001 and revised the proposal to reflect comments from the Intellectual Property Constituency before the proposal reached the form eventually adopted by the ICANN board.
Prior to adopting the VGRS proposal, the Board solicited the advice of its General Council, held a public forum on the topic of the WLS at its 19 May 2002 Bucharest meeting, and requested and received a collaborative report on the proposal from the Names Council and DNSO Constituency in addition to creating an electronic forum for the public to discuss the topic.
Most of the 856 comments posted to the forum.icann.org bulletin board on the VGRS wait-list period expressed opposition to VeriSign's proposal. As discussed in the methods section, we separated the comments by their origin: 144 comments were posted by the 1% of users who posted the most frequently and the remaining 99% of commentators posted the remaining 708 comments. We randomly sampled 252 of those 708 comments and, upon a careful review, discovered 23 comments in support of the WLS Proposal and 205 opposed. We can be reasonably sure that almost 90% of the public posting to forum.icann.org's WLS bulleting board opposed the proposal.
Many of those who posted public comments on this topic argued that implementing the WLS would have one of the following problems: it would negatively affect competition in the secondary Domain Name market; encourage monopolistic business practices; and/or eliminate those businesses already offering wait-listing services to interested customers. Registrars with an economic stake in stopping the implementation of the VGRS joined this chorus of dissension and encouraged its supporters to make their voices heard on the forum.icann.org page. In a 17 April 2002 analysis for the Board, ICANN's General Counsel also mentioned that legitimate business interests could be endangered by implementing the proposed WLS but did not endorse either acceptance or rejection of the proposal.
At the 19 May 2002 Bucharest meeting, representatives of the Names Council, registrars, and the Intellectual Property (IP) Constituency among others listened to VGRS present its wait-list service and subsequently asked questions and expressed their opinions of the proposal. While the IP Constituency had no objections to the WLS, the majority of the reactions at the Public Forum touched on many of the objections vocalized in the Internet forum and were similarly opposed to the implementation of a VGRS Wait-List Service.
On 24 July 2002, the Names Council formally adopted a report prepared by the DNSO Transfers Task Force recommending that the VGRS WLS proposal be rejected and, in the event that it were to be accepted over the objection of the Names Council, that it be amended in a number of ways. While acknowledging the Names Council and DNSO's recommendation against implementation and the comments made in both the public forum and on forum.icann.org, the Board decided that ICANN should support the WLS lest it deny Internet users a potentially beneficial service. The Board voted 11-2-2 to amend its bylaws and entered into negotiations with VGRS regarding the implementation of its proposed wait-list service.
Unlike the case of certain TLD applications, this proposal by VGRS for a waiting list service had extensive, and often quite carefully reasoned, public commentary in opposition to its passage. In addition to the public commentary, the Names Council recommended to the Board that they reject the proposal; likewise, a staff member urged caution. Despite this opposition, the Board moved ahead on the proposal. This example suggests that the ICANN Board has taken certain important action on at least one issue that is arguably policy-related or regulatory in nature without the strong support of the Internet community at large, regardless of how such input is most effectively conveyed to the Board.
Third, we reviewed ICANN's reassignment of the .org domain from VeriSign to a new operator . On October 14, 2002, the ICANN Board selected a proposal from the Internet Society as the basis for the successor operator of the .org domain. On this topic, the ICANN Board sought input from four recognized groups  who contributed to the staff evaluation of eleven applicants . The Internet Society received strong reviews from these recognized groups. The public commentary, which was relatively sparse, favored an applicant that failed, IMS/ISC. IMS/ISC did not meet the technical requirements set forth by Gartner. The Internet Society received more positive than negative comments, but at a less favorable ratio than IMS/ISC. In this instance, the public comments may have played a minor role in the Board's choice of Internet Society as successor for the .org domain, but the modest level of public participation meant that the technical advice of Gartner likely had the largest impact on the Board's decision.
During a 14 October 2002 meeting, the ICANN Board of Directors selected a proposal by the Internet Society as the basis for the successor operator of the .org domain. The resolution was adopted by an 11-1-3 vote.
The Board received eleven applications from organizations interested in replacing VeriSign as the operator of the .org TLD. Examinations were conducted on each of the candidates by four evaluators: (1) the Gartner Technical Evaluation Team; (2) the NCDNHC Usage Evaluation Team; (3) ICANN's General Counsel; and (4) the Academic CIO Evaluation Team.
The four evaluation groups were each charged with examining the candidates on criteria for which the evaluators had expertise. No other formal groups are publicly cited as factors in the staff evaluation. While public comments were received during the evaluation process from important organizations (e.g. VeriSign), these comments tended to focus on the evaluation process itself and not evaluate the candidates for the .org reassignment.
The contributions of the four recognized groups in the final staff evaluation are summarized in the table below:
The ICANN staff appeared to rely primarily on the Gartner evaluations to restrict the field to a set of "clearly competent" proposals. From this set, the NCDNHC evaluations were then used to rank the leading proposals. It is not clear how the General Counsel's evaluations were used beyond providing "corroboration of acceptability" on certain relevant criteria.
The selection of the Internet Society as the successor .org operator did not reflect the distribution of public commentary (Internet Society had 2 opposing, 4 supporting), though the public commentary with respect to this issue was relatively sparse. The IMS/ISC application received the greatest public support (0 opposing, 18 supporting) but did not meet the technical requirements the Gartner group recommended. Public commentary did appear to play a role in informing the NCDNHC group on evaluation of certain criteria. As with the selection of new TLDs, public comments received on forum.icann.org did not play a primary role in the selection process, but were acknowledged during the decision making process. The public comments informed the ICANN Board but were not influential enough to change the evaluator's recommendations or the Board's selection of the Internet Society as the new operator of the .org top-level domain.
Finally, we considered the proposal to reform ICANN's structure in 2002. This reform process began in earnest with then-president Stuart Lynn's Case for Reform of ICANN in February, 2002 and continued through the publication of a Blueprint for Reform, which summarizes the findings of the Evolution and Reform Committee (ERC). While there was no single board vote in favor of or against a single element of reform, ICANN has moved roughly in the direction of Lynn's reform proposal, as refined by the ERC. Throughout this process of reforming ICANN, the public commentary has been almost universally opposed to Lynn's reforms and the reforms that the Board has implemented, such as the move away from an At Large Membership and toward a less-direct method of public participation in choosing Board members. Much of this opposition expressed anger at the notion that ICANN was abandoning all elements of representative democracy and further empowering states. The At Large Advisory Committee, chaired by Esther Dyson, which was designed to represent the user community, submitted a final report that proposed recommendations not dissimilar from the Lynn proposals. Outside the work of Ms. Dyson and the ALAC, there is no evidence that the Board has taken the Internet user community's input into consideration throughout this ongoing reform process.
When then-president of ICANN Stuart Lynn issued a report entitled "ICANN — Case for Reform" on 24 February 2002, the massive reform process that would reorganize ICANN and change its relationships with governments and the public at large began in earnest. ICANN's new structure, as proposed, would give governments a more central role in ICANN and greatly diminish the direct role of the Internet user community in the ICANN decision-making process.
In 2002, the ICANN board passed a series of resolutions at its quarterly meetings related to ICANN's evolution and reform. Before the 14 March 2002 Accra meeting and soon after it, forums were created on forum.icann.org to gauge the public's opinion of Lynn's proposal. On both of these forums the Internet user community began to criticize the reform effort.
On the forum created for comments on The President's Reform proposal, there was universal opposition (46 opposing, 0 supporting). There were 85 opposing comments and 9 supporting comments posted in response to the EVR's call for general public comment. The comments on these boards frequently claimed that governments do not adequately represent the Internet user community and, therefore, users did not want governments to represent them within ICANN structure. Lynn's vision of a public/private model for ICANN called for a board of 15 people, five seats of which were reserved for government appointments, and the possibility of some form of At Large advisory organization. In addition to these structural reforms, Lynn also proposed introducing an ombudsman and a manager of public participation to increase the transparency and accountability of ICANN. Those who posted to forum.icann.org were almost universally opposed to then-President Lynn's proposed reform.
At the 14 March 2002 Accra Meeting, the Board officially moved forward with the reform process, while promising that the newly-created Committee on Restructuring would pay attention to the public forum comments. These public comments are cited as the impetus for altering the reform agenda to insure cultural and geographic diversity, an At-Large Advisory Committee, a greater role for the public in the nominating committee, and a renewed promise for a transparent bottom-up consensus model of decision-making. These goals reflect many of the desires of the Internet public, but fall far short of the ideals expressed by the commentators on forum.icann.org.
Before the 31 August, 2002 Shanghai meeting, at which the reformed by-laws were adopted, the Board created the Assistance Group on Proposed Implementation of an At Large Advisory Committee (ALAC Assistance Group) to help them implement an At Large Advisory Committee. The highly-respected Esther Dyson and other leaders of the Internet public were appointed to the ALAC Assistance Group and, on 19 August 2002 they submitted a "Report on the Evolution and Reform Committee" to the Board. Essentially, the report worked within the limits set out in the Blueprint for Reform adopted earlier by the Board. The Report fleshed out the design for a hierarchically structured ALAC and asked for a voting or non-voting liaison to the board, five regional delegates on the Nominating Committee that selects Board members, and liaisons to the other support organizations.
As per tradition, a public forum was created on forum.icann.org soliciting commentary on the ALAC Assistance Group's report, but the bulletin board received very little input. Four comments were posted. One was an official opinion from the IP Constituency and was discarded. One suggested that ICANN adopt a structure similar to the IETF, while the other two expressed dissatisfaction with the impotency and lack of voice with which the Internet user community had been left.
When the Board voted to implement reform, the new bylaws reflected the recommendations outlined in the Report developed by the ALAC Assistance Group, but did little to allay the concerns of the Internet public still posting to forum.icann.org. The opinions expressed on forum.icann.org, while belonging to only a vocal few, had no discernible effect on the reform process.
The bottom line of this initial review is that public commentary in the public online forums in favor of, or against, a certain proposal before the ICANN Board of Directors does not correlate strongly to an outcome either for or against that proposal. In several instances that we reviewed, the Board in fact voted or acted against the position adopted by the majority of users who commented through the public forums that we reviewed. Other types of input to the Board, such as the recommendation of a relevant Supporting Organization or of a hired technical reviewer, correlate more strongly, though not perfectly, to a certain outcome by vote of the Board of Directors. And on certain important issues, few members of the Internet user community bothered to express their perspectives publicly. As such, the online public forums that we reviewed did not play a significant role in achieving a representative and open decision-making process at ICANN.
These initial findings, along with other evidence and in light of the extensive academic literature on this topic, lead us to three conclusions:
First, the failure of ICANN's experimentation in achieving a new means of "open" and "representative" corporate governance shines further light upon the need for an overhaul of its governance structure. This overhaul should move ICANN away from its semidemocratic past and toward a structure better suited to the entity's narrow technical management purpose - and such reform ought not, necessarily, include great public participation through the means tried, relatively unsuccessfully, to date, such as posting to public message boards, mail lists and direct election of a subset of the Board, each of which modes appears to have sharp limitations. Other institutions seeking input from the global Internet user community should pay heed to the shortcomings of these mechanisms as well.
Second, ICANN should clarify the way in which users can involve themselves in the decision-making process for managing the domain name system. If ICANN remains serious about striving toward a "representative" and "open" decision-making process, which may or may not be the right decision, then it should make plain to members of the global Internet user community how they can participate meaningfully. The current organizational structure of ICANN is so complex, and the modes of interaction with the institution so numerous, that users cannot be blamed for failure to understand how they are meant to participate. ICANN's current public materials continue to point users toward the online public forums and do not make clear enough the best way to participate in the decision-making process.
Third, to the extent that we seek new means of governing the technical architecture of the Internet in an increasingly networked world, we ought to look beyond ICANN, which may never have been the right place for such experimentation given its limited technical mandate.
Our initial analysis represents only a part of what might be done to analyze data related to public participation at ICANN. A more thorough study would ideally: 1) consider more extensively the multiple inputs that the Board reviewed, in particular the extent to which individuals have been able meaningfully to participate through the Supporting Organizations; 2) delve more deeply into issues other than the four discussed here; 3) consider further channels of participation in more depth, in particular participation through the complex SO structure, and compare these channels against one another more explicitly; 4) study trends in participation over time; and 5) seek to segment the Internet user community in a more sophisticated fashion than we have done here.
While a law student and as an affiliate of the Berkman Center, John Palfrey attended the ICANN meetings in Melbourne, Australia, and Montevideo, Uruguay, in 2001, to assist in the provision of various public participation mechanisms, including the webcast, IRC, and related real-time services. He was one of several Berkman Center-related students on these webcast teams. He was paid a stipend for his work at these two meetings.