|
October 31, 1999 Los Angeles, California ICANN and the Public Interest: Pressing Issues I. Welcome - Diane Cabell A. Briefing materials online at http://cyber.law.harvard.edu/icann/workshops/la II. David Johnson - Wilmer, Cutler & Pickering. I represent Network Solutions but I"m not speaking for them today. III. Joe Sims - Jones, Day. No good deed goes unpunished. IV. Bill Drake - Exec. Dir of Western Justice Center Foundation. We create, test and learn from new approaches to resolving conflict. V. The NSI/ICANN agreement builds consensus right in. A. NSI's obligation is to abide by consensus policies -- which means no requirement to abide by other kinds of policies. B. Sims: Purpose of agreement to 1. end a war 2. find a way for companies like NSI to be comfortable with ICANN processes by showing process's limits, calculability. C. Sims: NSI had claim to own assets. Disagreement about that. Contract is way to eliminate the conflict. NSI gets benefit of assets for certain period, with certain parameters, after which all agree no claim. D. Sims: Consensus policies term is way to frame resolution of future, yet-to-be-defined disagreements (unlike those knowable ones of C, above). E. Bill Drake: So, define consensus. F. Close look at definitions in NSI/ICANN agreement. G. DJ: Consensus required is not a simple democracy majority wins. Instead, there are additional requirements not about (1) appeals to reason or (2) political representation. Instead, it's about getting a document that shows there's been adequate outreach and discussion among stakeholders, such that remaining opposition can be overridden. 1. The report is not just a document prepared by someone working alone has to convince stakeholders, IRP that there's broad consensus and that dissent can reasonably be disregarded. Report has to be convincing. H. Nesson: But what if there's some policy NSI doesn't want to comply with? 1. Sims: There's a review mechanism specified in the document (1.b) Independent Review Panel. IRP decides whether or not the board has met the criteria specified in the agreements. Decision to be based only on the existing record to encourage stakeholders to get their comments in early. I. Nesson: 1.c says further review if NSI disputes the decision of the IRP. 1. Johnson: Consistent with traditional processes in administrative law. Tries to anticipate the rare unexpected situation in which the document not consistent with the requirements stated. 2. Sims: Wanted to translate the generally-accepted concepts into a process that's actually workable. Hope that the process described is in fact consistent with the principles. J. Nesson: Suppose there's a policy NSI really doesn't want to comply with. Suppose money is no concern - could NSI resist? 1. Drake: Perhaps not so long. Notice and comment process leads to a determination then IRP if necessary. 2. Johnson: Don't know what the test of when you've got consensus will ultimately be. Will depend on intensity of support for proposal, the nature and intensity of opposition, the need for resolution of the question. In certain areas, ICANN has no power to assure that there's centralization in decision-making. Won't take a vote and require a particular level of support, but there has to be generally agreement. 3. Sims: No one wants ICANN to be a global regulator. Is to be a process, a forum for the creation of consensus, not a regulatory body. K. Drake: Who has power in this process? 1. Johnson: It's a significant challenge to reach out to constituencies that don't have as effective organization. Working Group design of DNSO requires NC to be creative in bringing in qualified WG members. If a necessary group hasn't participated in the decision-making process, the resulting decision is invalid. L. Nesson: ICANN is to function by consensus. If there's no consensus, ICANN can't set policy? But what if there's no consensus on particular issues. 1. Sims: That's the principle on which ICANN is based. This method is unproven - would be nice if it worked, but that's not certain. M. Nesson: How vulnerable is this process to a well-funded organization out to establish dis-consensus? 1. Drake: Less vulnerable than a system with voting or one based on expert analysis. The test is whether there actually is widespread agreement to a given policy. Judgment is based on the merits not political seat-claiming or alleged experts. Forces the conversation to take place among all those involved. 2. Nesson: But what about worries of faked grassroots campaign, hired guns, etc.? * Johnson: Would be extremely difficult because the process is open. Anyone who puts out a sham claim that particular stakeholders support a policy when that's not the case will find that the stakeholders in question will come forward and state their actual position. Any process can fail, but this seems preferable to the alternatives. * Sims: It's not consensus that's the bottom line, it's compliance with the agreement, and the agreement is highly specific about what has to happen. N. Sims: Definition of consensus. Dictionary says unanimity, but that's certainly not what's intended. O. Drake: ICANN and NSI are struggling towards a workable decision-making process. Can't anticipate all the questions that may arise, but common sense likely to be helpful. Journey to consensus is worth the detour of this process. P. Sims: Membership is harder - less intuition re structure of decision-making process. VI. Questions from Audience A. Phil Osborn: There was a disagreement between NSI and ICANN re ultimate appeals process. Binding arbitration? Or court? Many international questions thought better resolved by arbitration? 1. Sims: Either party can go to court to enforce the contract. Except re definition of consensus policy, for which both parties must agree. B. Abril i Abril: Concerns about procedure by which agreement was reached. Not taking a decision is a decision too. Invisibility of US Government is disturbing. C. Nesson: Would these agreements pass the consensus test? 1. Johnson: No. This, like contracts with registrars, but maybe that's OK because community will have to show consensus nonetheless by agreeing to be bound by the contracts. D. Leslie Harris: Dispute re scope of what's being decided. Some say ICANN will consider a narrow set of technology questions, while others are concerned about public policy issues. 1. Johnson: There's ambiguity in the middle - everyone agrees on certain things ICANN definitely will address and certain things that ICANN will not address. 2. Sims: No/little dispute within ICANN itself. E. Eric Menge (US SBA): Concern about outreach clause. Filed comments to this effect on Wednesday. Will this topic be left to WG-E? Will NSI accept a recommendation from WG-E? 1. Johnson: How to do outreach is a serious question for WG-E and other parts of ICANN. Shouldn't be as big a problem if ICANN stays true to its intended focus. F. Outreach isn't so hard if there's time and dedication. (WG-E meeting tomorrow.) 1. Johnson: Hard to keep people working on the relatively boring topics of the intended work. 2. Sims: Remember that ICANN is less than a year old. Haven't had intensity of effort of ICANN staff, public, etc. focused on outreach because there has been other pressing work also. G. Forman (Register.com): Agreed that consensus has been hard. But it's hard to agree to trust us when agreements of such importance are not getting the expected consensus treatment. Consensus should start now, not later. H. There are specialists in decision-making, consensus, etc. Some additional challenges - doing meetings worldwide, for example. But should build on established expertise. 1. Drake: Need to bring in minority viewpoints. 2. Sims: Lots of time & effort has been dedicated to issues other than outreach. Feels confident that outreach will be effective with effort. I. Berman (CDT): Want to assure that ICANN does not make social policy. 1. Sims: Agreed. J. Coates: Many members of DNSO are not in agreement. Can't say there's general consensus when there isn't. VII. Nesson: Feel like I know it when I see it, but trying to measure it makes it go away. VIII. Cabell: Introduction of tutorials. IX. Touton: New Registry/Registrar Agreements and Licenses A. Presented [PowerPoint Presentation] B. Questions - None X. Uniform Dispute Resolution Policy - Rodin, Froomkin (via videoconference), Kornfeld A. Rodin - proposed UDRP began with NSI policy with input from WIPO 1. Wanted a policy that could be used by all registrars, large and small, to govern all kinds of domain name disputes (not just cybersquatting), wanted resolution of cybersquatting disputes to be fast and inexpensive, wanted to balance interests of all stakeholders. 2. Presented policy in Santiago. Board approved it then and instructed ICANN staff to organize a small drafting committee. 3. Policy to be incorporated by reference into domain registration agreement. 4. Mandatory administrative proceedings for bad-faith registrations: Domain name is identical or confusingly similar to a trademark; domain name holder has no legitimate rights or interests in domain name; holder registered and uses the domain name in bad faith. * Bad faith: Domain name registered or acquired primarily to sell it to trademark holder at a profit; a pattern of such conduct by registrant; domain primarily registered for that purpose; misdirection (i.e. misspellings of trademark). * Rebuttals to allegation of bad faith: Domain name has been used (or preparations have been made for use of it). Or registrant is or has been known by that name. Or making fair use without intention to tarnish the mark. 5. Providers: Several under consideration. Complainants choose providers. Panel can be either one or three people. Either complainant or domain name holder opts for three-person panel. If complainant picks large panel, complaint pays whole cost. If respondent picks large panel, costs split. 6. Mutual jurisdiction registrar's principal place of business, or domain name holder's address (as of date complaint filed). 7. Reverse domain name hijacking - if a domain name holder can demonstrate legitimate rights in the name, complainant would have to go to court to continue to fight. 8. Board has approved the text from the drafting committee. Intend to have most registrars use the policy by end of 1999, NSI by January 2000. B. Froomkin 1. Two roles - ICANN skeptic and member of small drafting committee, then as disputes.org 2. Not an expert on antitrust, but concerns there. 3. General philosophy: This is intended to be a lightweight policy. Only for clear-cut cases. Other cases will still go to court. 4. Notable aspects of the policy: * These provisions more appropriate for business than individuals. Don't take into account people who go away for extended vacations, don't check email so frequently, etc. Notice starts when sent, not with any indication of receipt. * Doesn't penalize reverse domain name hijacking. Some suggest fines. Perhaps a necessary evil would end up worse off than now if we tried to solve the problem. But should be frank about what's been accomplished. * Tarnishment has different meanings in different countries. * No incentive for NSI to notify reregistrants in other countries. 5. As a participant in disputes.org, a potential provider under these rules: * From point of view of potential providers, these rules are tough to comply with. Not designed with providers in mind. Different in that lumping together similar cases (to reduce costs) is unlikely -times of cases are too unpredictable. Unlikely that multiple cases will share the same three-person panels. Not clear how many providers will be willing to serve with each other. Not sure if web-based process will work, so cost savings have not materialized. C. Kornfeld: Bill passed extremely quickly. Now House and Senate bills have to be reconciled. Differ from ICANN agreement because damages can be assessed in addition to domain names seized. Only applies to people whose domain names are registered by US registrars. 1. Criticism: Would undermine ICANN's extensive study of the issue. Favors trademark holders. Expands tarnishment definition. Doesn't account for possible legitimate interests of individuals in trademarked names. Global issue; US responds. ICANN trying to give global answer. D. Questions: 1. Sternlight: Examples of domain name hijacking? * Prince case (ultimately settled). * Many disputes never get to court because when a small individual gets a cease and desist letter, he often doesn't know how to respond and just gives in. 2. Complainant has too much power. Alternative suggestion of truly binding arbitration. * Rodin: Don't want to replace the ability to go to court. Just want to resolve certain kinds of disputes more quickly & efficiently. 3. Sheppard (NC): Trademark is regional and gTLDs are global. Hypothetical: Czech tractor company called McDonalds. Would this UDRP protect such a registrant. * Froomkin: Would the arbitration be considered binding in non-US jurisdictions anyway? But solving this might increase complexity & cost. * Rodin: Parties will need to go to court for hard cases. 4. Mueller: Complaint chooses provider. What if some providers not (thought to be) fair? Doesn't this tarnish the fairness of the process? * Froomkin: Think all arbitrators are likely to be impartial. Price competition intended to make the process cheaper, faster, etc. * Rodin: Public should review decisions of the panels to make sure they're being fair. Trying to keep fees low, recognizing that domain name holders may not be able to pay. 5. Field: From NameSecure, an accredited registrar with its own policy because they don't like NSI's. Names to be registered on a first-come-first-serve basis. Registrars shouldn't get involved; will just do whatever courts say. Have not yet been sued. Question: Who decides whether a dispute falls under the guidelines? * Rodin: Rules specify what a complaint has to state, and provider decides whether or not complaint states a case. XI. Emerging Privacy Issues - Davidson, CDT A. Davidson: Concern whether ICANN should deal with these questions. These issues may go beyond technical coordination to policy. 1. Introduction of panelists Shapiro, Delmers. 2. Domain name registrants may not realize that registration results in publication of certain private information (address, telephone number, etc.) a problem for individuals, some small businesses, non-commercial organizations expressing unpopular minority viewpoints. Can't just lie about this information may forfeit domain name if caught. 3. Should decisions like whether or not to have a .XXX be made by ICANN? Seems more like policy than technical coordination. B. Shapiro (MPAA): WHOIS is helpful for many purposes pursuing email spam, etc. only if info is accurate and comprehensive. WHOIS can be consumer protective. 1. Service providers who meet certain requirements, especially that responding to notice that infringement is occurring on their system, gain certain protections. This can't work if WHOIS isn't fully accurate. 2. Does WHOIS need to protect anonymity and privacy? Pirates, pedophiles, child pornographers and other bad actors are the ones who benefit from WHOIS that is not comprehensive/accurate. Policing benefits outweigh privacy issues. Anonymity can be reached without opt out. Example: Balance between consumer protection and privacy. Consumer calls to report that he looked at Disney Films to buy something but discovered that it was a pirate. Monolith and Anonymizer provide certain level of privacy without sheltering lawbreakers because we can still find out who service provider is and get bad actor shut down without necessarily identifying bad actor 3. Is it ICANN's responsibility? Governments are national whereas Internet is international. This arises in registrar accreditation agreements and therefore ICANN seems to be the best and most reasonable choice C. Delmers (European Commission) 1. We're not opposed to cooperative agreements in principle. We like new version of how registrar data is handled 2. EU directive on privacy of information is applicable: 1) collection of private data on registrants and 2) transfer of data outside of EU because ICANN is global system/mechanism * EU directive has Art. 25 that aims to ensure adequate level of protection for data collected in EU and disseminated globally 3. EU directive offers solutions: * Self-regulation protocols (codes of conduct, safe harbors) that are being discussed with US and are achievable within ICANN policy 4. Concerns about lack of precision/details that ICANN policy offers. Waiting for response from ICANN to these concerns. We are faced with policy that is pretty vague, especially the purpose of data collection. There are technical requirements but what about other reasons as well. * Need for ICANN to specify in written policy why data is collected, why and how it will be exchanged, what rights are linked to data for those it concerns. Specification will help as a limitation. D. Davidson: Is there a middle ground here? Not everyone who seeks to keep their info out of WHOIS is a bad actor, so can we make distinctions according to what type of actor wants his info kept private 1. Shapiro: Seems that ICANN is the authority to determine this and can clarify for European Commission (EC). The info we are talking about is not sensitive under EU because it is simply address, name. E. Nesson: Issue is anonymous speech. What about people who want to have voice on Internet without having to self-identify even in most basic ways? 1. Davidson: That's only part of it. It's also pseudonymous speech. 2. Shapiro: Copyright pirates shouldn't be entitled to anonymity 3. Nesson: What if only names that no one could care about were given out in ways that allowed anonymity? 4. Davidson: Some numbers will always be better than others. 5. Shapiro: We're not fighting over the domain name itself but over what's being done at the domain name. We'd like to catch the direct infringer but we'd be satisfied with getting the service provider to block/remove the website (digital millennium copyright act) and then we can get name of infringer using subpoena 6. Davidson: But I see that you are talking about setting up a special area that allows anonymity (ie a domain). Will we allow people to create such domains, and for what allowed purposes? 7. Shapiro: I think some technologies/ areas for this already exist. But all the pirates will end up here, too F. McKowski: 1. Delmers: That is up to ICANN to decide whether that is allowed or not. But notion of consent of customer to dissemination is not expanded in new version sufficiently. G. Osborne: I'm troubled by fact that differences in laws/mores of different jurisdictions will be applied to registrants. Question: what about people who are already registered but whose information has previously been kept private, and who may face repercussions under new requirements that information be public. 1. Some ccTLDs currently allow anonymous registration so if this standard applied to them, it would be a change H. Crispin: There is currently no difficulty in getting access to Internet services anonymously, and whether there is ability to get second level domain name anonymously is really not a big deal 1. Davidson: So one answer is to get a third level domain name because then you have total WHOIS anonymity. Low barriers to entry and wide distribution protects individuals seeking anonymity. Is the danger that this will become a centralizing force because now everyone will have to go through someone else to be anonymous? 2. Cabell: Geocities, etc, could stop allowing anonymity -what if there were no such providers left? 3. Crispin: If terrorist wants to get to someone who is anonymous, they will go up the chain until they can identify someone and they'll bomb the identifiable party 4. Davidson: Accountability is a big deal. Access to domain name system is access to Internet in a real way. CDA was opposed on basis that Internet is extremely decentralized and there is not a scarcity of access points so should not be controlled in same way as e.g. broadcast medium, where there is scarcity of bandwidth 5. Crispin: How is ICANN the bottleneck? I. Carre: Engineering point. Accident v. design. WHOIS is accident rather than necessity. ICANN making policy decisions about accountability is early when in fact there will be (and are) other technologies that will be better vectors for these issues TLD Competition XII. Introductions of Panelists A. Mockapetris - invented DNS. Can speak to what the founding fathers intended. B. Hill - Netnames, a registrar. C. Sheppard - Business Constituency of DNSO. D. Postrel - Economist. E. Mueller - Professor specializing in telecommunications policy, involved in WG-C. F. Anonymous member of ISP constituency. G. Crispin - Chair of PAB of gTLD-MoU. H. Porteneuve - Advisor to AFNIC, DNSO secretariat. XIII. Do we agree that new gTLDs should happen soon? A. Sheppard: depends on process for choosing which gTLDs and how. XIV. Constraints on adding domains to root? A. Mockapetris: Start slow and ramp up. Adding twenty would be OK from a technical perspective. B. Mueller: ccTLDs have been added quite quickly. Might provide insight into dynamics of adding domains. XV. Models A. Internet Stability Model -- Crispin 1. Places most emphasis on stability. Hedge bets against registry failure - standardization of data format so that records could be transferred from one registry to another if necessary. 2. Zittrain: Emphasis on standardization in code? Where would standard come from? * ICANN. Any database of the size of .com would take significant time to copy. But for smaller databases, data copying shouldn't be a major concern. B. Mueller: The Root is a resource. Want to define how to share the use of the resource. Need to maintain stability. Add, say, six initially chosen from applications from whomever is interested. Because stated intention is to add many more later, not so important which six are chosen initially. C. Hill: Netnames might consider applying to be a registry. D. Mystery Guest: Some ISPs are entering the registrar business to provide additional services to their customers. But most are outsourcing. ISPs may not be so interested in being registries. E. Postrel: Skeptical of importance. Having a different naming system is perhaps less important than everyone thinks. The part at the front of the domain name is important, but the particular ending is not. Do need some sort of editing process to help choose what sites are worth going to, but that's a service provided by directories. F. Sheppard: Differentiation might or might not help users find things. It's confusing to have .film, .movies, .audiovisual, etc. Better just to have one. 1. Zittrain: This is ICANN's responsibility to say that we'll have one gTLD per category, and to define a set of categories? * Yes, with sensible consultation with others. G. Mystery Guest: SIC codes already sort businesses. EC has something similar. But choosing a coding system will take more effort, time, etc. than letting directory systems take over. H. Mueller: Initially thought classifications would make sense for gTLDs. But gTLDs need to remain fixed over time, while classifications need to change at times. I. Mockapetris: Perhaps this needn't be a zero-sum game. Try several different policies for assigning gTLDs. Needn't force particular classification systems. For example, country needn't be the most significant bits of a name system that's why we have gTLDs and not just ccTLDs. XVI. Zittrain: What kind of diversity do we need in ccTLDs? XVII. Zittrain: Is Mockapetris's idea of diversity inconsistent with your proposal? A. No. Proposal has a concept of sponsored gTLDs suggested by a particular company for a particular purpose / business plan. B. Zittrain: How exactly to do it? 1. New gTLD would have a purpose, requirements for registration, who would enforce requirements. 2. Would accept proposals for for-profit gTLDs although expect that WGs wouldn't accept them. C. Porteneuve: What about worry of different requirements for particular classifications (medicine, for example) in different places? Who would be allowed to register in .medicine? D. Crispin: Such worries presuppose particular charters. These questions would be asked in the WG process, and if no resolution could be reached, then that working group wouldn't be a good choice. E. Remote comment read by Zittrain from Dennis Schaefer. In short, does a trademark holder have rights on their trademark in all domains? F. Crispin: Nothing gives exclusive right to be ham operator in my proposal 1. Mueller: But they have to go through public process, which will make it political inevitably and will result in having strikes against them if they meet with success 2. Zittrain: Is there a way for ICANN to yank franchise if registrar allows registration by not qualified entities? 3. Crispin: Short answer is that each sponsor takes responsibility under the contract they sign with ICANN. Ex: .museum registry would be contract between ICANN and museum group that would specify prices, qualifications, etc sponsor is not a registry but is a database company 4. Mueller: Under my proposal, museum group would say we want to run .museum registry, would be allowed, then if other museum group wants to do the same, can do so easily. Crispin's proposal, Mueller says, would not allow for ease of registry for other museum group but prefers existing group G. Engineer (Romit): What is the hype for the g in gTLD? Why not let the name space maturity reflect the maturity of the group? No difference between . extension and slash extension 1. Zittrain: so the fight is misplaced? 2. Maybe groups should be visibly splintered just as much as they are in reality, just up a level 3. Mockapetris: There is some value in the .com area, aside from being proof of resources to be there. But let's try a few different things, so we try 3 or 4 best ideas and let the best one survive. 4. Mystery: Vixie made a comment that has escaped this panel: DNS is not a directory service for marketing. Why do people want these other things rather than just going after slash? It doesn't fit on a business card (ie isn't sufficiently sexy to go after a slash). 5. Sheppard: Consumers need to be able to find what they are looking for, and that should be the basis of whatever solution is found and that problem exists for both commercial and noncommercial entitites 6. Crispin: DNS is not recursive 7. Zittrain: Is that practice or necessity? 8. Crispin: Practice. But DNS is not uniform at every level H. Menge (SBA): DNS wasn't designed to be marketing tool, and isn't best choice anyway. If top level domain is tied too closely to what business you're in, are you then in serious trouble if you change your business model? Who decides whether you've strayed from your category? I. : ccTLD is de facto monopoly. To facilitate competition and better service, need to separate functions ccTLDs registrars and ministries and define laws and responsibilities. Need to encourage new entry and additional registrars 1. Mueller: Separating registrars and registries doesn't actually help competitive issues because it makes the business of registering a name or updating contact information less efficient without adding competition way to do that is to add domain names 2. Portaneuve: What do you mean about ccTLD competition? Registrars can compete within it, but .fr is only competing with .net and other .*s. ccTLDs will compete with other ccTLDs in terms of whether the regulatory scheme is favorable J. Lalande: Extensions don't intrinsically mean anything. The public is being led to believe that they do mean something. Simplified domains.com K. Amadeu I Abril: Competition in domain name registration market is only on supply side. They will fight for customers but not really against other registries. .com's will not shift to .biz, so we need to make this natural monopoly as gentle as possible. But setting prices and getting customers within a domain can be done by market. System needs to be self-healing and with little need for oversight. L. Cookey: Internationalization of DNS. ASCI code and Latin alphabet have so far dominated. What is view of panel of internationalization of domain space for non-Latin alphabet names? 1. They like it, it's already happening. M. So what six would you start with? 1. Portaneuve: I don't have six names. Until we can be sure that economy will not be harmed, I am not ready. 2. Crispin: Just core TLDs. 3. Mystery: None, or three letter nonsense extensions with no meanings (Vixie proposal) 4. Mueller: I'm not playing because my position is that ICANN should not be choosing. It should be first-come, first-served six. 5. Postrel: I also think anyone who wants to pick one should be able to do so. (.god) 6. Sheppard: Only six? We've learned nothing from .com monopoly. We could start with .thinkagain, .confused, .ripoff 7. Hill: I'm not giving any top level names either. Intuitiveness matters for users, and if it gets hard to use we've lost whole point of having top level at all. 8. Mockapetris: Bruler wants a domain name to go to Native Americans. I would like opt in or opt out phone book. I think we have no problem with picking the first ones. 9. Cabell: .sex, .drugs, .rock&roll Openness and Transparency XVIII. Nesson: How to balance need for openness in order to create trust with intimacy to create trust among Board for effective decision-making? A. How much of a director's life is now open? (Email?) XIX. Introduction of panelists A. Frankel (by videoconference) B. Lorne C. Crocker D. Schwing XX. Nesson: What can be kept private? A. Schwing: Attorney-client, purchasing, etc. B. Frankel: When board talking only with each other and not with outside world. When negotiating to be able to change positions more easily. Afterwards still have to disclose decisions, arguments, etc. 1. Nesson: So closed meetings OK so long as motivations are explained afterwards? C. Nesson: Does email before a physical meeting need to be open? 1. Lorne: Don't favor establishing transparency as a goal, rather translucence. Want to provide an opportunity to say stupid things without the world seeing and knowing as much. Wouldn't want every such stupid comment to be public, but enough material should be public so that people understand what's going on. Maybe release more information after the fact if ultimate decision seems questionable. But presumption should be against the sort of detailed disclosure ( interrogation ) that's suggested by some. D. Crocker: It's impractical to think everything will or should be out in the open. Even when larger portions of the decision-making process are open, suspicion remains. (So openness in and of itself won't solve a lack of trust?) E. Frankel: ICANN is unique in terms of its model. Like a government when it comes to making policy decisions. Explanations of motivations for decisions will produce more trust than simple openness of meetings. F. Nesson: ICANN Board Member Greg Crew is in the audience. How do you feel about your ICANN life being public? 1. A number of board members were opposed to open board meetings initially (New York meeting last fall) in order to get more work done. These nine board members have formed a strong working group with only a few meetings and teleconferences; closed board meetings helped. G. Schwing: In most states, two board members cannot meet for dinner the night before a meeting. So have to be worried about going to the same church or being on the same PTA. 1. Nesson: So what should ICANN do? Don't want courts setting aside ICANN decisions eight months later! Do want board to be able to work together. But want openness. H. Crocker: Need to allow open input on the worker level. Sometimes working groups inefficient perhaps in part due to openness, but deadlines can help move things along. I. Cukier ("cold-called - by Nesson): Openness would kill reporters - business, for reporters make their living by opening up information that's otherwise closed. Right now ICANN is getting off pretty easily - a few reporters from trade press, plus some mainstream press (New York Times). Later ICANN won't be able to get away with, for example, having closed meetings then changing to openness when subject to Congressional oversight. Perhaps assumption should be that everything is open except that specially closed. 1. Nesson: Email? * I take what I do seriously, and having email be open helps me do a better job. FOIA adds heat to officials by reminding them that there's accountability to what they do. J. Nesson: Is email ever to be open? 1. Frankel: Sometimes. FOIA provides some insight into how to proceed. K. Crew: Anyone should know that their email is not private and should be careful not to say anything that would expose them to litigation. My email is already open because it goes to the whole ICANN board L. Osborn: What about possibility of conflict of interest of ICANN directors as they become more representative of special interest groups? Suppose ICANN were to set up oversight watchdog committee or opened it up to public that rewarded whistleblowers. We need for all to have stake in regulating regulators. 1. Crocker: Need counterbalancing force so not spending all time defending selves - need opposite bounty to discourage false claims 2. Lorne: There are some circumstances in which we are entitled to know quite a bit about those with power - a different kind of openness 3. Frankel: Some danger in this kind of approach; gives ability to forward private agendas using public platform. Should retired judges or some other objective group have this authority? M. Amata: Only when public body does its business in private do issues arise, not simply relationships or private rendezvous. Today's arguments were rehash of traditional government arguments against openness. Constituents have a right to know how decisions were made. ICANN guidelines give no guidance for when meetings should be closed. Shouldn't all be open unless specific guidelines allow closure? 1. Crocker: Visible but respected is missing in conversation about making all open. We don't have sense of proportion about reasons for openness and subjects for openness. Tone has been: everything should be open and then defend anything closed but perhaps should be opposite N. : Closed email exploder lists? 1. Crew: Sometimes they serve a purpose, sometimes makes sense to start private before going public. O. Crew: Dynamic is going toward more openness except board meetings, which haven't been open because many directors believe that closed meetings are more efficient for trust-building and getting things done. Critical issues are now policy, which are not board purview in their formulation so openness values will be preserved as the issues are publicly framed for the board's vote. ICANN board openness is growing, although the obligation to do as constituents expect is not the same for a corporate board as for public elected officials. Membership XXI. Announcement: Asia Pacific TLD (APTLD) group meets Friday at 8:30 - 6:00. XXII. Thanks to meeting sponsor Markle Foundation and to co-sponsors CDT and CPSR. XXIII. Moderator: Leslie Harris XXIV. Panel A. Greg Crew, ICANN Board & MAC B. Postrel, Reason Magazine C. Aizu, APNIC & MAC D. Graham, O Melveny & Mysers E. Hill, Center for Voting and Democracy F. Amato, Citizen Advocacy Center G. Cabell, MAC XXV. Introduction by Crew A. Last major action of initial board membership to choose new elected directors to replace existing At-Large directors. B. Phasing by council and (new and old) board members to give testing of process C. Key issues remaining are details for process/supporting systems to enable very large global registration of members and voting to occur 1. No money for this, either D. Not unreasonable that at-large constituency will have to pay for own costs just like Supporting Organizations do E. By-laws as written contain specific requirements and prohibitions. However, silence is permissive rather than exclusive. 1. Harris: Lawyers' relief act? XXVI. Tension between ICANN's view of its role (narrow, technical, managerial) and need to gather global constituency - how are these issues likely to engage large numbers of people? How to maintain balance? A. Postrel: It is a temptation, but these members have lives beyond Internet. Technical or business interest will drive activism unless ICANN's role is expanded. Slates will emerge because ordinary people will not be engaged sufficiently to vote on non-brand (party) basis 1. Could be religious, nationalist, other special interest groups that could find issues sufficiently important to participate in large numbers and possibly dominate even if not real world majority B. Harris: Why define according to verifiable address and not domain name owners only? 1. Cabell: Technical protocol issues affect larger numbers of users than simply domain name owners C. Amato: Direct vote is important because issues to be decided will grow beyond technical ones. Four guiding principles require ICANN to attract large constituency rather than discourage membership. Right now, package makes membership unappealing 1. Cabell: Why vote for president? Represntation is important value, even if unwilling to pay for it 2. Harris: But there are bounds on what electoral college can do, but not on what Council could do. Example: Council could elect selves to Board. 3. Aizu: Looking at how many users there are in the world, which is group at-large members will come from * Looking at graphs of users * Different regions have very disparate penetration rates * How long will it take for the regions where penetration is low to catch up? * Are issues really universal, ie how do we address diversity of situations? 4. Harris: So an even smaller group faces these issues? 5. Postrel: Membership in at-large council is designed to compensate against regions with high penetration in order to protect equities among geographic regions. Country code is anachronism. Netizens versus national citizens. 6. Harris: Architecture is destiny. Was structure created to engender certain results? 7. Hill: Basic message is that method of election (cumulative elections, choice voting using proportional representation) is operating system of organization that defines who can get elected to govern * Proportional representation will give broadest and most diverse representation on board * Half of board is Support Organizations. They are by nature better organized and financed. Other half will not have axes along which to unify, so you have power imbalance. Market driven SO's will be able to unify due to shared incentives * Reduction of representation is inevitable given two-tiered system * At-large requires 5000 or more for coherent election. SO's don't have to have a threshold. You may never have a valid election if low voter turnout. 8. Graham: In real world, this will not result in good board of directors. What are you really trying to accomplish? A quality board representing the geographic and other interests of the Internet community, or just let anyone living have a shot at a seat? Balkanism is inherent in what we have. D. Harris: What was the vision of the ICANN originators? Checks & balances? 1. Crew: ICANN is in existence to create a mechanism whereby the Internet community can develop policy for the orderly development of the Internet. Need to balance needs of community using (and that will use) the Internet against the knowledge and abilities of those who profit using the Internet. 2. Hill: Structure in place will not achieve what is intended because electoral college mode suggests suspicion that masses should be involved but not too closely. Second round, with majority on council, could put all of their candidates on board. Narrowing of points of view that can be represented on board * So council allows capture of board because of majority threshold for board election 3. Amato: I think system was designed to keep individual users out of decision-making process, and it does so. Membership is discouraged and then size of 5000 required. Membership can be dissolved, but you don't have similar requirements for SO's 4. Graham: Presupposes that election means you will be an effective director. Experience shows that learning to be a director takes time and the problems discussed here are exactly why boards are usually filled by identifying groups to be represented and then finding best people to do that. Corporation will be harmed by election of bad directors. Do we care about constituencies or the corporation at the end of the day? You don't know what you'll get at the end of the day 5. Berman: Bylaws have been improved. 6. Cabell: I wanted to see them improved because I have clients who are domain holders who had no say in how their rights would be treated. The non-technical, non-hegemony voice was not being represented. Non-Californians were being governed without representation. E. Harris: How will we conduct this election? Do we imitate US national elections? How do we prevent abuses and protect values of representative governance? 1. Schaefer: I am disappointed that ICANN has allowed capture of Board by commercial interests 2. Hill: Stakeholders on SO side could get seats on membership side and then capture whole government 3. Postrel: Don't try to prevent slates; let them happen so we know where everyone stands and what their agendas are 4. Cabell: Organizations are supposed to be in domain name organizations. Domain name holders can't be individual members on membership side but must join domain name organization. I dislike fact that I have to join something to get a voice 5. Postrel: Slate mailers are result. Parties are better. Openness in slate positions, funding, are best. 6. Aizu: How is this practicable in global scale? Enforcement and implementation are problematic. F. So how do we deal with enforcement? How do we identify the circumstances of capture? 1. Graham: This is very much outside the usual method, and that's the problem. There are solutions to these problems, but not in context of a non-profit CA corporation. Wrong vehicle. Nomination process and direct democracy would be preferable. Nomination by current board or blue-chip committee. 2. Sims: First draft of pre-ICANN bylaws proposed four SO's. Fourth (we don't now have) was industry users. Rationale was that it was to be a technical, administrative organization so those in that realm should have majority of votes. However, lots of other people use Internet and should get voice. Got from there to where we are, a messy place to be, because of a strong feeling by one part of stakeholder group that we needed more direct involvement and influence by user community. This argument, of organizations vs. users, has been going on for a long time. Current board members have feeling of obligation to address this argument and have selected a crummy device to do so because it is the best of alternatives to address problems. * Personal (non-Board) view is that structure is wrong. Representative democracy is wrong vehicle. Process for input should be using organizations that represent users of various types but in aggregate form. 3. Hill: Corporate vs. democratic methods have been aired. Key issue is whether democracy makes sense here. Look at non-democratic WTO that makes enormously influential decisions. Representative government seems important in this context 4. Nesson: As Berkman Director, not moderator. We got involved in membership issue for ICANN from beginning because we thought it was the most vital. Original conception from Postel and Sims was beautiful. Biggest border in cyberspace is between proprietary and public domains. Explosion on proprietary side. ICANN has task of running fundamental property system in the space. Postel wanted to pass on a balanced board. How to reify a system of choosing nine people who truly embrace the mission of supporting the public domain? Inability to be sure of authentication allows ease of capture. * What about a different worldview, a change of structure. Right now it's choose 9 on one side, 9 on other, and they choose president to get 19. When in this legal reality I ask what entity or individual I trust, looking at the for- and non-profit entities, that axis strikes me as one possible division. What if we allowed the for-profits and the non-profits each to choose 9. The at-large could choose president. - President's powers are not for me to define. 5. Graham: It looks like any of this can still be changed, but if you get too far down path you can't 6. Postrel: It's not clear to me that limiting the at-large voting to certified non-profits (definition varies by jurisdiction, too) would work. I am also not sure that it addresses the problems that led to this discussion. Individuals still want a voice and they will still look for a label. 7. Graham: Piece that appears to me to be lacking is that the debate is always about to whom we will give votes but there is nothing on how to select the very best directors to achieve the vision we have described. That's the core for success and discussion has missed it. 8. Amato: Very few hours relative to the others have been put in regarding the membership. It's a very complicated issue and focus has been on SO's and their creation. Structure now in place has been put in without requisite groundwork 9. Crispin: Zittrain example: Man of the Century contest had Attaturk as winner for every category because Turk patriots made this their mission and no one else got as motivated. Capture is serious risk 10. Postrel: Everyone worries about for-profits capturing, but recognize that capture happens with Scientologists and Ayn Rand fans, too. It's a risk with any extremely motivated groups 11. Osborn: No drive to encourage or educate membership. 12. Crew: We don't have system in place for membership registration. 13. Osborn: What is ICANN, what is Internet, what about those who aren't there yet but would like to be and should get a voice that will help them get there. Not everyone has the same stake but all votes so far appear to be equal. Stake in property seems to be driving process now. Why not form Internet trust that people can buy shares in and domain name purchase gives shares and then run it as profit-making body 14. Hill: If you have decided that this should be representative rather than corporate, re-visit 9 and 9 seats issue. Did SO's get seats serendipitously? Why not diminish their seats and use proportional system for rest? Not necessarily geographic representation. Single transferable choice voting. 15. Graham: Keep it simple. Get board to articulate what they think the rest of the board needs to look like to satisfy the constituencies not yet represented. Decide exactly what the slate must look like, take nominations but make clear that decision-making power is in board. Must satisfy smell test. Right commitment to important constituencies could make this happen. 16. Berman: Half of the job was done by Postel. Now the other piece is what to do with the not-organized: the user community. What possibility is to create an interim board for that part of the Internet for a short period of time to make recommendations on how to institutionalize correct procedures. 17. Chong: Beyond next June, supposing that all goes well, what are the rights and obligations of the at-large members? Vote, but what else? Are they watchdogs? Or can they join organizations according to interests and stakes? They may have to pay money, and they can vote, we've told them. Can't stop at elections * Amato: Extremely difficult to make membership attractive when no power to remove directors, bring derivative action, amend bylaws, or do other acts that you can do under CA law. * Graham: People join organizations all the time without a thought for the other rights/powers. * Aizu: Visits to 5 Asian cities. Negative approach is so far most successful: if you don't join, you get shut out. 18. Zittrain: Tough problem. My brain hurts. As I look at the solution, I see that ICANN wants to do it right, but also knows it needs to get it done to protect actions from accusations of non-representational abuse of power. No one is really defending current set-up. No incentive to vote for thrice-removed-from-power representatives. If everyone hates it, does that make it a good compromise? I don't think so. Speaking as someone who had a role in the creation of this system, I think we need to challenge it as our final answer. Is what we have really better than the blank page we had before? We're getting new and helpful voices coming in. Delay on membership may be a bad thing, but it may be a lesser evil if we can try the Constitution without waiting for membership principles. * Postrel: We could create a slate of people for each position, present them as choices, and then let those elected from that figure out next steps * Crew: Board came into this with mission to establish membership for ICANN. Membership at-large entitles one to same things as membership in SO's. However, if we can't get at-large members to be as interested as SO members, it will fail. We can't be sure here today that we won't see membership interest. Why not give it a try before we start over. We let the SO members pick their directors according to their own needs, and we should do the same for the at-large. Electorates get the board they deserve * Aizu: Campaign for outreach has not happened because we have no money to do it. Why should only one US foundation be the funding source for global at-large membership * Hill: If you advertise this organization as the power over the domain name root server, you will get attention. Make it a representative body with recognition that it will be dominated by SO's with occasional input from at-large memberships 19. Harris: What's the worst that can happen if we follow our current course? Can we establish procedures that allow us to undo anything we find untenable? * Hill: Capture by commercial interests is not undo-able. * Amato: If ICANN becomes captured by commercial interests, efforts to remove it will intesify and that will be a check because ICANN will lose its power * Crew: ICANN will only continue to exist if it satisfies its constituency. Capture will nullify it. * Aizu: Capture of ICANN has happened: by US government. CONTACT INFORMATION For additional technical information, please contact: Ben Edelman and John Wilbanks |