Site Map

Please note:

You are viewing archival ICANN material. Links and information may be outdated or incorrect. Visit ICANN's main website for current information.

ICANN Meetings in Lisbon Portugal

Transcript - GNSO New Top Level Domains (TLDs) Public Forum

26 March 2007

Note: Although transcript output is largely accurate, in some cases it is incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the session, but should not be treated as an authoritative record.

>>BRUCE TONKIN: I'd like to get started. All those who don't want to be part of this session, please step out and have your conversations outside.

The purpose of this session is to present the current status of the new gTLD policy development work.

The agenda is roughly to cover some material as to why we should have new gTLDs, which I've heard that question raised a few times recently.

Then we'll talk a bit about some further work we've done on the application process, string checks, sorts of criteria for applicants, our thoughts around an objection process and dispute resolution processes, and new gTLD operator obligations.

Then Kurt Pritz from the ICANN staff will present some initial thinking from the staff on some of the issues this poses for them and how they'd go about implementing this policy.

One of the things I'm conscious of is that we, as in the new gTLD committee, have been working on this particular work probably for about a year and a half. But many of us have actually been working on this for more like seven or eight years. And what I want to do is give a little bit of perspective on that seven or eight-year perspective to put this work in context. Because there are a lot of new people who have joined ICANN or started to participate in ICANN meetings and I think they probably don't have some of the history.

So just at a very simple level, if you think about postal addresses, basically, they've evolved from being entirely human-based, in other words, someone writes something down on a piece of paper, gives it to someone else and says go and deliver that to this location. They've been human friendly, human readable. Only recently have postal systems begun to use automation and tending to translate these human friendly, human readable addresses to something electronic. In Australia, they use a unique eight-digit code which covers all the addresses in Australia. And more commonly, people understand post codes and country codes as also being something that's fairly automatable from a postal delivery system.

We've started with something that's human readable and now we've designed something that is machine readable.

Internet addresses have gone in the opposite direction. We actually started with something that was machine friendly. And so the ICANN Web site address is actually 192.0.34.163. You can actually go to that address if you actually use your computer and go to that address, you'll get the ICANN Web site.

But that's certainly not something that's easy to remember.

But now, if you actually look at those two forms, from a computing point of view, they're actually equivalent. They're just basically processing those by looking where the dots are and they're processing that information.

So the domain name system, the original goal of the domain name system -- this is quoted word for word -- the primary goal is a consistent name space which will be used for referring to resources. Those resources can be anything, but they're typically computers. In order to avoid the problems caused by ad hoc encodings, name should not be required to contain network identifiers, addresses, routes, or similar information as part of a name.

To give you a feel for resources, there is absolutely nothing wrong with this telephone having an Internet address and having a domain name. So domain names don't have to be Web sites, they don't have to be e-mail. And, in fact, there are actually coke machines around that have Internet addresses. So they're not necessarily related to Web sites or e-mail addresses even. It's just a location of a resource that's connected to the Internet.

The other key thing here is that it's not restricted to a single application. A lot of people now associate the Internet with the World Wide Web. The World Wide Web is only one of many applications on the Internet.

Now, in designing this name system, the idea was that it would be hierarchical and have a tree structure, and also it's both machine readable or machine understandable, and also human readable and human understandable. At a very simple level, we're starting the tree at a dot, then we at a single level can have a dot A, dot B, or dot C, then we can have a A dot A, -- a second level, and a third level, A dot A dot A, that's just hierarchy. Then we have choices between a deep or shallow hierarchy. So a deep hierarchy has many levels. And, typically, you have short names at each level, but you have many levels. So I could have A dot A dot A or B dot A dot A. They're two separate names, A and B, and the rest of it is indicating the levels. Or we can have a very shallow hierarchy with a few levels. But we tend to have longer names needed at each level.

It's been interesting to see the evolution of the way domain names are used. Originally, it was envisaged probably that we'd be somewhere in between these two things, probably a medium hierarchy, and a lot of earlier domain names were based around educational computers, so you might have a computer name that's something like Bruce dot electrical engineering dot engineeringfaculty dot monashuniversity dot education dot Australia. That's a pretty deep hierarchy. That was a common use of domain names was labeling computers like that. But more recently people have said they don't want to type all this stuff and put all these dots. They'd rather just have Bruce@monash.com.

The previously two slides were giving you options, a deep or shallow hierarchy. But at the end of the day, from a computing point of view, it's pretty much equivalent.

So the actual choice that was made in the domain name hierarchy prior to 2000 -- and this is from RFC 1591 -- was that we would use these two-letter country codes, which are actually borrowed, I believe, as postal codes. That's from the ISO-3166. In Australia, when I look at CH for Switzerland, that's not intuitive for me. That doesn't mean much as a human based in Australia. That's not how I would have thought Switzerland would have been abbreviated. That's just an example of something -- it might make sense in a postal system but isn't especially useful for humans perhaps globally.

Then it had these generic ones. And it was interesting to reflect back on the standard that I've often thought this organization was intended for noncommercial organizations. It wasn't. It was the miscellaneous. It's a very typical engineering approach here saying, I can think of commercial, I can think of network operators and I'll just have something for everything else, which we'll call org. The definition of org is actually miscellaneous. Then it has a suggestion there saying, well, you could actually, if you didn't fit into any of these other categories, E.G., nongovernment organizations, you go here was sort of the advice and the standard. So that's where we were prior to 2000.

And then the requirements to operate one of these TLDs, whether it's a ccTLD or a -- or one of these com, net, or others, has also been sort of specified in a standard. And it's worth reflecting on those, because when you come to look at how we've designed the system, they pretty much reflect these requirements.

So -- and I've numbered them in the order that they were in the original RFC. So one is where you first have to designate a manager to supervise the domain name space. And this is, I think, the point that Bertrand raised earlier. It says these designated authorities are trustees for the delegated domain and have a duty to serve the community.

So the sense of it is that it is a public resource here, and that being allocated one of these top-level domains, you have a responsibility to serve the community in doing so.

It also says that the manager must be equitable to all groups in the domain that requests those domain names. So, in other words, if your domain happens to be .AU for Australia and it's intended for people in Australia, you need to be equitable to all of those people in Australia that are applying for those names.

Another point here is that significantly interested parties in the domain should agree that the designated manager is the appropriate party.

And I guess this is where a lot of the redelegation work happens, where there's a dispute as to who should run the ccTLD in some cases.

And then -- and it's pretty much judged on that basis, do the people that are significantly affected or involved in that TLD agree that you should run it.

Also says that the manager must do a satisfactory job of operating the DNS service for the domain.

And we've incorporated that in our technical requirements.

And then the last point here is referring to transfers of management from one domain to another. So this is -- this is an existing document RFC which was originally developed around the -- around the original top-level names.

Now, when we consider those categories, I mentioned we had these country codes, and then we had com, net, com for commercial, net for network operators, and org for miscellaneous. Now actually look at the fact that today we don't have conformance to those categories.

But look at TV, which is the code allocated to the country of Tuvalu. Nowadays it's commonly associated with television or video. And if you go to www.tv, which is the Web site that's telling you about dot TV, there's no mention that I could see of Tuvalu at all. It was all about you should get your domain name here because it's a great place for television and entertainment.

So that has very little to do with the TUVALU postal code.

Dot com is meant to be for commercial entities, but, actually, it's widely used for Internet users, including natural persons, legal entities, profit, nonprofit, or just those that can't manage their organization.

It's driven -- the growth in dot com, interestingly, has been driven by the low cost to the registrant as well as the ease of registration. And what's tended to happen is because some of the country codes originally had some strict rules about who was able to apply for those, and dot com didn't and it was often cheaper than those country codes, a lot of people -- this is effectively competition has happened, the market force has meant that a lot of people have moved to dot com, and dot com has internationally massive volume of registrations.

Then what do the country code names do? Some of them started to duplicate the top level. In Australia, we have dot com, dot AU, and dot net, dot AU, so we try to have a similar hierarchy. But now many ccTLDs are choosing more flat hierarchies. And that seems to have been an ongoing trend to have a less deep hierarchy and to have a relatively flat hierarchy with only two or three levels as opposed to, you know, five or six.

So that shouldn't say motley, it should say "mostly registrations" are occurring at the second level for country codes.

Interestingly, some country hierarchies have now been duplicated and in dot com. You can go and register within UK dot com as well as AU.com. And that's generally driven by entrepreneurs that sort of say, well, because com dot au has got all these rules, that's okay, you can have UK.com. So they're effectively creating their own hierarchy within dot com. We're starting to see that we're actually getting whole name spaces starting to appear within dot com, because there's not a mechanism for having new names at the top.

The other thing to be mindful of, thinking about domain names, is that as a global trend, especially for people in this room, there's widespread access to convenient travel and to communication technologies.

And what this means, I think, is that natural persons no longer necessarily associate themselves with countries anymore. Often people travel a lot, plus they're in communication with people that are at other -- you know, that are across border. So people routinely, if they have a particular user group of interest, are communicating with people that are no longer in their local village or in their local town or even in their local country. They're spread all over the place.

And we see people now with multiple citizenship and multiple languages. And if we were going back several hundred years ago, that would be very rare, because most people probably could only go as far as you could walk. And most people don't want to walk a hundred miles, so that would be about as far as they ever got.

So, you know, that's changed the nature of the environment.

And part of that global trend is that companies and organizations are now global.

I work for a global company. We have offices in numerous countries around the world.

So companies are starting to operate across borders. So if you look at our current categorization, we're saying we've got these top-level domains based on country and then these things based on commercial and network operator. It doesn't really reflect how people are using the Internet today.

So by background as to what's happened at ICANN, ICANN's really been working on this problem since the beginning of ICANN. And the initial ICANN working group on new gTLDs began in June of 1999. And I went back and I read the report this morning. And I saw that the arguments for and against new gTLDs are pretty much the same as the arguments that I see currently before us today. This is eight years later.

All I need to do is change the date on that, and you wouldn't be able to tell the difference.

And so I thought I'd try and summarize them very briefly. Obviously, there's lots of views for and against.

I think the for argument is saying there should be increased choice, diversity, competition, all arguments that are put forward. Providing an opportunity for global communities to have their own hierarchy.

And against, unnecessary, you can just register at the second level in the existing hierarchy.

And that's probably a little bit like saying it's unnecessary to have your own house, because you can just have a room in a hotel or, in fact, you just have a -- sit on a plane, which is where I sleep most of the time.

Or -- but, you know, obviously, that's an argument either way.

The other major argument against has been the requirement for defensive registrations to protect brand.

And that's very much the nature of the things that I just mentioned earlier, that, originally, if you look at the trademark environment, you have produced a product, you made this product in your local village or whatever, and you are competing against the other person that made the same thing, whether it's making shoes or something, and so you might stick your mark on the shoe and you want to show that this is a shoe that I've made, and you want your mark to be different from the other person. But it really only had relevance to that local area.

And then we've seen, as companies have gotten bigger, they're saying, okay, we're selling these shoes all the way around the world, so we want to protect the brand in each country. So they've got trademarks in each country they operate in.

And what we're seeing, a carryon of that behavior, I've got it protected in dot EU, protected in dot com just as a matter of course, now you're adding these new gTLDs, because I'm going to have to spend more money to protect it.

That would also argue against your creating a new country for the same reasons. Because if you add a new country, they're going to have to obviously protect their trademark in that country. So the background, then, behind the group or the work so far is that in 1999, despite the sort of for and against arguments, a decision was made to introduce new gTLDs at that time -- this was 1999 -- and begin a test bed with six to ten TLDs.

And there was a call for applications. And in that first round, there was quite a bit of diversity. It was deliberately done that way, that there was more applications than TLDs that were released. They tried to get a bit of diversity, so they had some that were fairly generic and open like dot biz and dot info. And then they had others that were a lot more specific, like dot museum, and dot coop. And, you know, these TLDs in terms of people's awareness of them have varying degrees of success. As you'd expect with trying anything new.

Then in 2004, there was another limited round. And in that round, we saw more diversity again. If I look at three there, .cat, dot travel, dot Asia, they're all completely different. And they don't even match any common taxonomy or hierarchy here. If I was going to look at a hierarchy based on travel, then you might say, is that a recreation? So I should have dot travel and dot golf and things like that. And then .cat, referring to the Catalan community, so, therefore, if I'm going to have these names that relate to language, then, you know, maybe I have, you know, dot Spanish, et cetera, et cetera.

Dot Asia is a geographic term, typically. So we have a combination of a particular industry; we have a language group; and we have a geographic group. So totally different in those three areas.

So one of the things that people have talked about over the years is should we have a structured taxonomy, because that's how it started. We started off with a list of postal codes and then com, net, and org. People said rather than just these random names like what we just had, travel, cat, and Asia in these different areas, surely we're smart enough to come up with a taxonomy and agree up-front what the set of names should be.

But when you really start thinking about it, that's not easy. So what are we talking about here? The purpose of the Internet, is it primarily for commerce? So my taxonomy should be finance, manufacturing, all the traditional industry codes that you see in government statistics.

Or is it about the users of the Internet? So what, should I have the dot Mac users, the dot PC users, and the dot other users, because I'm not -- I ran out of ideas at that point, similar to the com, net, org approach.

Or are we actually talking about a community of interest that people want to exchange recipes, so I have the dot fish eaters, the dot meat eaters, and those that don't eat.

If you do this, you can see, how do I actually choose this taxonomy. And I'm only choosing things that make sense to me as coming from a particular cultural area.

If we are truly looking at the diversity of the Internet, it really -- the view of the GNSO -- and this was a decision that was probably made in I think around 2004 or thereabouts that we looked at this, basically, we came to the conclusion that we didn't believe it was possible to come up with a taxonomy that really met the needs of Internet users.

So essentially we have gone with allowing communities to self-select, using, I guess, a market-based approach.

So I've just kind of mapped out, I guess, some of the history and some of the arguments for and against.

And then I've just tried to select one that I think is perhaps the most compelling in the context of I guess the growth of the Internet internationally and the worldwide trends in the way people are using the Internet.

So I've only just chosen one particular reason here, and that is to support the functional, geographic, and cultural diversity of the Internet by allowing globally distributed communities. I am using the term "globally distributed" in the sense that these communities are not located in a particular country, so it doesn't necessarily make sense to add them to dot AU or to add them to UK, because they are globally distributed, the opportunity to have their own hierarchy of names starting at the top.

And I think we need to accept that not all communities identify themselves with other countries or the existing broad categories of com, net, org.

And I think that's just something that needs to be accepted rather than debated.

So that's my attempt to try and answer that question. There is no one answer. And we ran a session here in Marina del Rey of this committee, and everyone on the committee had a different answer to this question. So I've just given one.

But there's no definite answer. There's obviously diversity of view.

So the new gTLD committee has basically accepted the outcomes of the 1999 work. It was one of our input documents. And then we've taken into account experience with introducing new gTLDs so far.

And we focused on lessons learned in doing that, and we focused on creating a process for introducing new gTLDs that we hope will work for the next hundred years. Obviously, we're not going to get it right, but that's the approach.

And we've used the ICANN mission and core values wherever possible to guide the work. Because at many stages in our development work people would come up with good ideas, and we would say that's a good idea, but is that really ICANN's role, and does it match ICANN's mission and core values.

The other thing that has come up in the lifetime of new gTLDs is the concept of Internationalized Domain Names and the standards around this didn't exist in 1999. So relatively new in this process.

And I think the committee supports the introduction of IDNs when technical testing is completed.

But we, as much as possible, again, treat them the same as any other gTLD from the process point of view.

And that's really the policy decision we've made.

We do recognize, however, that IDNs make implementation quite complex. And I'll give some examples of that as we go through our process.

But what we did do to try and understand some of those issues further was we formed an IDN working group and that group released its report last week, and on the weekend we went through that report and just tested whether our process would actually work for IDNs. And by and large, it does.

So I'll just pick up some of the things, because at the last ICANN meeting, I did go quite a few of these things, and I won't focus on the ones that we've discussed already but more focus on changes since Sao Paulo and some of the debates we have had around language.

Firstly, string criteria. Basically, we're saying it must not -- a new string that you add to the top level must not be confusingly similar to an existing top-level name. I think we probably spent a hundred hours on just those two words, confusingly similar. And I am not going to go into detail as to why, but at a very high level we are trying to choose words that are in common use in different legal frameworks around the world in similar circumstances. And we borrowed some of the approaches that have been used in trademark law. And our report sort of documents different excerpts, I suppose, from trademark offices and different trademark laws around the world. And this seems to be the most commonly used terminology, I guess.

Not cause technical instability is an obvious one.

Then the other one we are trying to understand what the legal rights of others might be. So at a high level we are saying a string must not infringe the existing legal rights of others, and if I put a full stop there, then the question is, well, which legal rights? There's probably a (inaudible) if you combine the number of countries by the number of laws, you are obviously looking at thousands of potential legal rights.

So what we have tried to do is focus at least our process on those rights that are recognized or enforceable under generally accepted and internationally recognized principles of law.

So we're really trying to sort of say that even though some of these principles may not be part of some international treaty, by and large, very similar laws exist in multiple countries. And therefore, those laws that protect the legal rights of others would be the ones that we would focus on.

To defer the work in this general area around what some of these rights might be and particularly how to handle this at the time of registration at the second level, we formed a protecting rights of others working group. And that's due to complete its work by May.

Another topic that is surprisingly complex is around what words you can't have. And so generally, we're saying at a policy level that you can't have a reserved word, and then the question is what's on the list of reserved words.

The kind of categories that exist at the second level currently, so for new gTLDs there's an appendix in their agreement that specifies a number of categories of words that are reserved at the second level. And a couple of those there are interestingly that IANA and ICANN names are reserved. In many cases, different -- slightly different wording in different agreements, but single and two-character labels are reserved, and then names with hyphens in the third and fourth character, and that's to allow the introduction of Internationalized Domain Names.

Again, this is quite a complex area, surprisingly. And we formed a working group to look at that, and it's got about a 150 page report on this topic.

And so we're trying to sort of extract from that report those areas where we think there's pretty common agreement that we use as part of this process. But obviously this reserved word concept is going to be ongoing.

And the feeling is that when the name is rejected for a reason, such as it seemed to be confusingly similar to an existing name, it would be put into a disputed list, and that's advertised. So a subsequent applicant can say, oh, okay, this is in a list of words that have been rejected. So it wouldn't make sense to spend your money on an application, and you will immediately get rejected.

The other controversial text or text again we probably spent many hundred hours on is trying to understand the nature, I guess, from a public policy point of view that strings must not be contrary to generally accepted legal norms relating to morality and public order.

And again, we're trying to sort of focus on those that are generally accepted rather than that might -- you know, be very specific to one country. That's doesn't mean that that one country still wouldn't be able to use its laws and the legal processes that exist, but we're trying to create a process where we would actually provide some assistance in terms of dispute resolution to resolve a dispute as opposed to using the courts of one specific country.

So that takes us to complaint and dispute resolution.

Generally, the idea is that the strings are posted, the applications are posted. The community is able to raise issues associated with where the strings match those criteria I've gone through.

The technical disputes, the intent is that will be resolved primarily within the ICANN structure, because it is an area within ICANN's core expertise.

But where possible, where the issues are outside of ICANN's core expertise, and giving some examples would be that one, I don't think ICANN is expert on that topic, and most likely the confusingly similar one as well, and rights of others.

So those are all examples of things that we, the committee, didn't think as something that ICANN should try and suddenly create another hundred staff to look at those things.

We're basically saying that there often, in many cases, are existing bodies that should be able to assist in resolving disputes on those issues.

And the key thing here is that the decisions will be made with reference to internationally recognized principles of law, rather than something we just make up as ICANN.

So that's some of the string issues.

Now, the applicant issues, the first two have already been presented in previous meetings, and that's essentially that they can demonstrate that they know how to run a technical infrastructure, and that they demonstrate that they have the organizational capability to fulfill all of their obligations. And a lot of discussion just prior to this meeting on the session about what happens when a registrar or something goes out of business, it's kind of related to this thing. When you're accrediting someone, it's not just the fact that they are able to run a request but can they fulfill requests for transfers and all the other consensus policy recommendations.

The new one that's been added since the last meeting, and this is, interestingly, fairly well matches the RFC 1591 in this area, but the way we've phrased it is if someone wants to apply for a particular TLD that there are obviously strongly interested parties, the criteria for not accepting it is that there must be no substantial opposition from among significant institutions of economic, geographic, cultural or language community for which that TLD string is intended to support.

The example we have been using is if someone applies for dot bank and certainly they are not a bank and have never actually run a bank and you find that most -- if the various banking institutions around the world saw it, they would complain. But we are not looking at a complaint from just one bank here. We are saying the banking industry would need to coordinate among themselves and be able to present to ICANN saying: We, the significant banks of the world, all believe banking this particular applicant is inappropriate to run that TLD.

So that some of these words like dot bank, dot police, et cetera, and having a mechanism for deciding or for objecting if someone applies for such a name.

String contention occurs when multiple valid applications for the same string or confusingly similar strings.

The way we propose to manage the process is to do it in rounds for two reasons. One is in terms of staff being able to manage all these applications at once but also to look at collisions between applications. And where people are either after the same string or a string that's confusingly similar.

So when that happens, the first approach is to encourage applicants to resolve this problem amongst themselves. There's probably three ways they could do that.

One, they meet together over -- at the bar at an ICANN meeting and sort it out. I haven't included soccer match is another method.

Mediation, where they would use a third party to help.

And finally, they could elect amongst themselves to use a binding arbitration process.

This is what we hope they do.

We also expect that, based on other activities of ICANN, not everyone will agree.

And so the other fall-back mechanisms is, I guess, two approaches. One is if there is a significant established institutions of economic, geographic, cultural, language community for which that TLD string is intended to support, they would use a comparative evaluation process, probably fairly similar to what we did for dot org and dot net. So we are basing saying we go through a process where you have to demonstrate that you are the best party to run that TLD. And it's a pretty expensive, heavyweight process here, in the hundreds of thousands of dollars for all parties.

Alternative, if it's something very generic, so someone wants dot red or dot blue or dot meat eaters or something, or dot fish eaters, they are not related to any particular established institutions anywhere, they are just a community of users, then most likely auction would be used as the simplest method to resolve the contention. And most likely the cheapest.

It's expected it's probably going to be cheaper for most cases than comparative evaluation for TLDs that are obviously not very valuable in their own right.

So the idea here is that for this additional work and the additional evaluation work involved that there be additional fees for applicants at that stage. So you pay a fee to apply for a TLD, and that goes through all the normal checks. And if there is no contention or no objections, then you get your TLD.

If, however, there are multiple parties going for the same TLD, then there would be an additional fee charged, assuming, of course, that they haven't used this option first.

Basically, we're saying if you are smart, you will do this first.

But if you are not, you get in here.

So I guess that's kind of an update on some of the work we have been doing.

And then in terms of our next steps, we feel that we're -- we've converged reasonably well on a process, and we're seek to go finalize those recommendations at a GNSO level by May, and produce the final final final report. One of the things we need to do is get better at the way we label documents, but we've had this thing called a final report that's been around for probably a year, but it does always have "draft" in front of it.

Then we would submit what's called the board report to the board in early June. Exam then available for the board to consider at its meeting in Puerto Rico.

So at that point, I'll hand over to Kurt, if he is around somewhere, to give us an update on the -- I guess the staff perspective. And one of the things the staff has been doing is trying to create some flow diagrams for how this process might look. And it's been very helpful for the committee to sort of look at taking the words we have produced, put it into a diagram and say, hey, that's not what I expected, and then we have gone back and done more work on the words.

But I think Kurt will be able to show us some of these diagrams, hopefully.

>>KURT PRITZ: Somebody tried to paid me a compliment earlier today, told me I could take simple concepts and make them complicated. I so hope I don't do that here.

I think this has been -- this effort to introduce new TLDs to develop the policy and then subsequently to implement that policy has represented a significant step for ICANN and especially for ICANN staff in working with the volunteers of the GNSO and sort of hand in glove in understanding their concerns from a policy development standpoint, testing them out from an implementation standpoint and feeding some information back that's been very well received, and having the policy development be iterative with some of the issues that are anticipated with the implementation of that policy. And I think it bodes well for future PDPs.

Understanding the keen interest in the community about the introduction of new TLDs, staff has done work in testing how the policy as it's being developed might be implemented, and I want to describe the work that's been done, some straw men implementation models for the policy as it stands right now, understanding it's rough and not certain.

But it's certainly to invite additional comment, because it seems that the more people who understand the difficulties and complexities with this can comment on it and help refine everybody's thinking.

So staff, in doing its work, has taken into account several sources of information. One is the council work to date. And, you know, among that is the draft final report that Bruce just referred to.

There's experience from the 2000 TLD round and experience from the 2003 TLD round. The net and org bidding processes. And then ICANN staff discussion points presented to the council on this date, and actually now on other dates. And also there's draft GAC principles that are being amended now, but we understand some of the thinking of the GAC.

In doing that, ICANN has developed implementation principles, so if you read the draft final report, there's policy principles for the development of this process and procedure -- policy process and procedure.

We also have implementation principles. Certainly chief among them is ICANN's security and stability mission.

But after that, that new gTLDs -- whoops. That's -- you know, I was working with Ross Rader last night and he added that functionality to my laptop which I thought was cool at the time.

[ Laughter ]

>>KURT PRITZ: But I have misused it several times today already.

New gTLDs, second most important will be deployed in a transparent, predictable and smooth-running process. That's Paul Twomey's words, "smoothly running process" so I wanted to capture chose.

The next is communication with the applicants ensures that if there's any communication with the applicant, the entire process be identified early on so that the applicant has early and full knowledge of the full extent of the application procedure.

So in that regard, ICANN will provide for frequent communications.

ICANN will also provide in implementation of this a former mechanism for applicants to settle conflict outside of the application process itself.

Often, independent evaluation panels will be used to make objective recommendations and provide advice, and that the board will be responsible for making many final decisions.

Application fees will be designed to cover costs, and that finally, technical and financial criteria will be objective, so that clear bars will be established in the RFP, but that experts can use their expertise to clarify expert -- expectations.

ICANN has started budgeting for this process. It's supposed to be self-funding but there are significant startup costs. We think, as we go through the process, you'll see the requirement, perhaps, for three, four, or five different independent panels that will be compensated in some form. Not, certainly, as full-time staff, but in an on-call basis so that those people can be called on as necessary and react in a timely manner, so that applications can be processed in a timely manner.

So as far as at least identifying where the costs will be, we see somewhat of additional staff levels as far as being able to man the organization. One staff -- or an FTE to oversee the engagement and the operation of the panels, a staff member to oversee the management of the process itself, and the processing of applications, and then an overall program manager to manage those efforts. And also administer to the new TLD program and the changes that are made to it.

Necessarily, other ICANN staff or outside consultants will be matrixed to the program, so inside and outside legal counsel and I.T. would be very important in this regard.

Another one-time effort will be to develop an RFP that will provide clear criteria to applicants so that they can -- so they can have an expectation of any issues with their application when they file it.

And then I talked about independent panel formation.

So I'm going to briefly describe the process as we see it.

It's essentially a three-stage process, so applications are received and reviewed for completeness. That's very -- a very brief process before they are posted, and the evaluation begins.

For those applications that not controversial or represent strings for which there will be no objection and that clearly meet technical and business criteria, there will be a preliminary evaluation stage.

Many applications, we think, at the close of that preliminary evaluation stage will be complete. So just to reiterate a little bit, if the application meets the objective technical and business criteria, or the string itself does not lead to technical instability or unwanted -- unexpected results in the DNS. So an example of that might be dot PDF or dot WWW, strings in themselves that by the very string that pose a security or stability issue. The string is not a reserved name or confusingly similar with an existing string or proposed string, or where there is no string contention, as Bruce defined it earlier. And then if there is no bona fide formal objection to the string, then staff sees a fairly brief process for processing applications.

However, if issues are raised in any of these cases, the application would go to an extended review process. And what we think is important is identifying for the applicant at that stage what that process will be, how much it will take -- how long it will take and about how much it will cost.

So if an application is problematic due to the nature of the string or string contention or some other issue, we identify for the applicant early on that there will be additional cost and time involved, and a good appreciation for how much that is. So that the applicant can make business decisions based on that.

So I want to take you to a flowchart that sort of....

Now, this flowchart is a work in progress. As a matter of fact, this morning we were trading e-mails back and forth and making changes to it.

On the other hand, it wasn't something that was put together, invented for this meeting. This chart has existed for many weeks, dating back to just about the first of the year. And we've had several discussions around it and refined it.

And on the left-hand side are the three phases of the application. I'm not completely signed up to the time frames involved here but they have been suggested and so they act as goals.

The preliminary part of the evaluation would be to determine that the application is complete. And then there's the initial or preliminary evaluation process.

So on the left-hand side, sort of, the applicant criteria would be to test whether there's technical or business qualifications are met for that application.

So at this stage in the initial evaluation there wouldn't be any iteration between the applicants and the evaluators. This would be if the evaluators found the application that meet the baseline criteria, it would pass that. If they found that it didn't it or had questions, it would be passed on to the extended evaluation period. The idea here is to provide clear criteria and a roadmap so applications can be dealt with in a fairly routine and quick manner and dealt with completely in the initial evaluation phase.

Bruce, in his talk, mentioned that if there was significant objections from an applicant under this applicant criteria, about the suitability for certain applicants to possess certain strings or have certain strings, that objection would be made next to this business and technical qualification.

So you can imagine maybe another box that contains that.

What staff had in mind when we put this box together is the public comment on the applications would include that objection. But this would be another part of the process.

And then on the right-hand side, we would test the string criteria and we would test it in two ways. One, that -- that were described earlier. One is the technical issues involved, whether the string itself would tend to harm the Internet. And, two, whether there were objections filed -- whether there are objections filed pertaining to this certain string, and I'll talk more about what those objections might be in a minute.

So taking the input from the technical and business evaluation and then whether there's any objections filed, staff would draft an initial evaluation report and obtain public comment on it. And then issue a final initial evaluation report, which is -- that term may need some work there.

The idea is that at this stage, the evaluation is over for those successful candidates, those candidates for which an objection is not made or those candidates whose application did not raise technical issues having to do with DNS stability or technical and business criteria associated with the application.

If there are issues raised during that time, then the application would go into an extended evaluation period. And so those extended evaluation periods would be intended to solve three issues. One is to resolve whether the application in fact met the technical, financial criteria that was posted as part of the application.

Two, whether it -- it did in fact result in potential technical harm to the DNS.

And, third, if there were objections to the string, that those objections would be referred to an external dispute resolution provider.

So imagine that there's a string to which an organization has an objection. That organization would formally object to the string. That dispute would be handed off to a dispute resolution provider. So the contention, then, would be between the objector and the applicant to settle. And the dispute resolution provider would, taking into account appropriate input from interested parties, would decide whether the string should be awarded to that applicant or not.

It's not a simple process and it's a long and expensive process. What we think is important is that ICANN identify for applicants early on in the process whether this long and expensive process is going to be included for this application, so that applicants can decide whether to continue or not.

All those things contribute to another report that will either recommend introduction of the TLD into the zone or not.

You notice on the left-hand side of this part of the flow, there's no time frame. I think the time frame is dependent upon the type of contention there is. And as the process is fleshed out, there will be different frames associated with different kinds of contentions -- with different kinds of issues.

So what kind of grounds can be used to object to a string?

We look to rely on posted policy, posted law that already exists.

So could the registration or the use of the proposed string violate legal rights under international law?

Second, could the registration or proposed use of the string be contrary to accepted legal norms relating to morality or public order?

Therm, does the string represent an established institution, NGOs, industry groups, and so on, as you can read.

And fourth is, is the string a geographic identifier?

So objections can be made only to limited -- for limited purposes and only by bona fide objectors.

If the objection's received, then the application would be sent to the dispute resolution provider.

Finally, I want to talk a little bit about string contention. So contention can be -- arise among and between competing applications requesting strings or generic words, or they can be competing applications requesting strings that could represent existing institutions or communities.

So parties will be afforded the opportunity to settle the string contention themselves throughout the process. But if at the conclusion of the process the string contention still exists and there's more than one applicant that has met the criteria and the objections have been resolved, the parties can elect to go into arbitration, or in the case where the string represents a community, a comparative evaluation process may be used by an independent panel. Or absent other methods, as described earlier, we might have an auction.

So that was a little bit complex, but I'm happy to take any questions.

It's a little bit -- yeah, I can....

There's also been quite a bit of development about a string contention resolution process. And so again, some work is done here.

Can I go through this, Ray, or do you want to make a comment?

>>RAY FASSETT: Okay, sure.

>>KURT PRITZ: It'll take me about five minutes.

So the first thing I want to call attention to is the box on this slide, asking whether this applies to IDNs. So I think an accepted principle is IDNs are going to be new gTLDs, just as new gTLDs are. So IDNs will have to comply with at least the requirements set out for new gTLDs.

Again, with string contention, if there's two or more applications for the same string, we have initial -- we have an initial resolution process and then a final resolution process.

But if the draft evaluation report, the initial evaluation report that was described in the earlier slide, identifies that there is string contention or that strings are confusingly similar, the competing applicants would be contacted, revising the contention, and a period of time will be given to the applicants to resolve their differences.

And if their differences are resolved, there will be a report to that effect. But if they're not, then the applications would go on to a more complex, expensive, and time-consuming process.

There's again three options here for resolving string contention. One is, the parties may agree to submit to binding arbitration. If they choose not to do that, then we have a couple different cases. One is the case where the string results to an existing institution. In that case, we would do a sort of comparative evaluation, and we would -- an independent evaluator would determine, among other things, which applicant best represented the interests of that community. But there will be more complex criteria than that.

And in that case, the applicants would be asked to submit additional material, go through a public comment period, and some sort of process that has to be well defined in order to determine which applicant is most suitable for that string.

In the case where there isn't a community represented, after a lot of discussion and the difficulties involved with trying to resolve string contention, there's close to a consensus that an auction may be used to resolve those differences.

So that's sort of the flow for string contention.

So that's -- Ray, did you want to -- or, Bruce, did you want to --

>>BRUCE TONKIN: No, fine with me, just open for questions.

>>RAY FASSETT: Is this on?

>>KURT PRITZ: Yeah.

>>RAY FASSETT: Thank you, Kurt. Actually, now I have two questions.

>>BRUCE TONKIN: Can you also just -- yeah, just state your name as well, just for the record.

>>RAY FASSETT: Ray FASSETT.

Just to be clear, if there is an objection, the -- the entire cost of the dispute resolution will be borne by the applicant? There will be no cost to the objector? Or is that decided yet?

>>KURT PRITZ: Yeah, there could be a filing fee for filing an objection. And so that would bear part of the cost of that.

But that's not -- I don't think the council is around that yet. And so certainly not -- staff is not.

>>RAY FASSETT: Okay, okay.

Second question, if you can go back to the flowchart.

>>KURT PRITZ: Which one?

>>RAY FASSETT: The contention one.

Yes, that one.

The upper left, go up -- down, down. Down a little bit.

>>KURT PRITZ: Sorry.

>>RAY FASSETT: If you can go -- just page down a little bit. The box that says "parties may agree to binding external arbitration," that box.

If -- just making up a hypothetical, if there are ten different parties and nine say, "Yes, we want to agree to a binding external arbitration," and one says "no," does that basically throw it to the right, then, as I'm looking at it?

>>KURT PRITZ: I would think so. And then the nine might decide to not submit to binding arbitration, given that of the nine, there would be one winner to contest with the one that didn't.

So I think there has to be some iteration so the parties can make decisions as we go along.

>>RAY FASSETT: When we say -- just to be clear, when we say parties may agree to binding, it has to be unanimous?

>>KURT PRITZ: Well, it's up to the parties. But I would think that's the way they would want to be.

>> ADRIAN KINDERIS:

Adrian Kinderis from Auth Registry.

Kurt, with respect to the technical evaluation, has ICANN put any thought to potentially streamlining the process with respect to the registry backend services and potentially doing some sort of pre-accreditation, like a registry accreditation -- excuse me -- for potential backend providers so that that was taken out of the evaluation period and potentially made it easier? So that, you know, would be something that you would have a potential list of accredited registries, and if their name was put to an application, then that criteria would not have to be reviewed?

>>KURT PRITZ: I think so. I think -- yeah, the really short answer is, yes. The longer answer is, the technical panel in the last round, you know, there were different issues with every application. And I'm not so sure there wasn't value added by raising those issues during the application process. But I think one of the hurdles for having a backend provider should be easily met by that sort of accreditation.

>> ADRIAN KINDERIS:

If I may, I would just for argument's sake, like to extend it out to not only gTLDs, existing gTLD registries, obviously, they'll work up stats, but this could be open to ccTLD registries or anybody else who wanted to get some sort of preapplication, preaccreditation?

>>KURT PRITZ: I agree.

>>RON ANDRUFF: Good morning, Ron Andruff, dot travel.

I want to thank both the GNSO and the committees. Excellent work done. And it's exciting to see a nice process and flow coming together.

I wondered if you might expand on the timeline, because it's tough to read the slide. You said that, obviously, no one's married to this time line now. But get a sense from application through completion, number one.

And number two, if there's a sidetracking where you have contention in the string, whether it's two parties or some other reason, has there been some timeline thought through on that? Because if you're going on a main track, you don't want to go on the sidetrack and sit waiting for staff to come back and deal with those issues when they're fast-tracking other issues, other TLDs. Has any thought been given to that yet?

>>KURT PRITZ: On this slide, it shows that the initial evaluation process might be complete in 85 days. I think the -- the issue there is the publication of applications to determine if there's going to be objections to it. And so there's been a lot of discussion wrapping around a way to communicate across all regions about the application process and then the results of the initial application round and the applicants and who they represent and what the strings are about.

So I'm not sure that that brief a time period affords feedback from those who might be interested, because certain organizations take longer to form an opinion and respond than others.

>>RON ANDRUFF: I'm inclined to agree. And, in fact, I would actually suggest we might just extend those timelines a little to let people do work in proper time.

So instead of trying to do it in 90 days, maybe we do it in 120 days, but let's make sure that we hold that timeline.

Then, sorry, my second question, Kurt, was, if there's a contention in the string, has there been some thought to that? We know there's a 90 day or 85-day plan, on the contention, is that still like an 85-day plan is your thinking now? Or is there --

>>KURT PRITZ: I think if -- I think if there's string contention between two similar or matching strings, not a lot of thought's been given to it, but I think it can be very well defined.

I think that for other sorts of objection, that, you know, there will be longer timelines that will have to be defined as part of the process, because the goal is to have applicants know as early as possible how long the process is going to take.

>>RON ANDRUFF: Excellent. Thank you again.

>>BRUCE TONKIN: I think the other comment I'd make there, Ron, is that the slide that Kurt's tried to put up there is as much as possible trying to encourage people to get into a process that finishes in let's call it 85 days, and recognize that, you know, if you -- here's the string criteria. This is the sort of stuff we're going to be looking at. Provided it meets all these things, it can be done in 90 days. And you're done.

Then it's sort of saying, if you don't, here are some of the processes we'd use. But I think the problem for staff right now is that apart from UDRP, where we have some experiences, as an example of a dispute resolution, we have an idea of time frames, a lot of these other ones haven't been tried and tested yet. So it's pretty hard to ask Kurt, once you get off the fast track, how long the long track is, because there are too many variables, I think.

>>RON ANDRUFF: Understood, understood. The only issue I was getting at is that the concern is amount of staff time. The trains have to keep running besides while we're doing that. I want to make sure everyone's keeping that in mind that we're allowing enough time that we can actually do these things and we don't find ourselves running over these days and then we're back into that situation where we didn't allow enough time. That's all.

>>BRUCE TONKIN: So --

>>RON ANDRUFF: As long as we've got a time frame, everybody is happy, I think. Just keep to it.

>>BRUCE TONKIN: Let me just comment again. This is personal view rather than staff view.

But I think the issue is how it scales in terms of both having sufficient funds available, so if you have a number of applications, obviously, you can get funds. But you've also got a training issue. I just -- it's more of a project management exercise; right?

Let's say you've had lots of applications and therefore -- and they each pay the application fee, so it's normally self-funding. It's not like ICANN can suddenly go and employ extra people, given the lead time to train those people.

So there are finite limits, I think, to, you know, how fast you can bring on board extra people to match demand. So that's going to be a bit of a lag, I think.

>>RON ANDRUFF: Very good. Thank you.

>>BECKY BURR: Becky Burr on behalf of myself, I promise.

Could you talk a little bit more about who is entitled to file an objection? I take it that the idea is not that anybody can file an objection.

>>BRUCE TONKIN: It depends on the objection.

Do you want to choose a category?

>>BECKY BURR: No, actually, I guess I don't. I mean, anybody who files an objection under any of those categories, whether --

>>BRUCE TONKIN: Let's -- well, if you're claiming a legal right, you would have to have standing to do that. So --

>>BECKY BURR: Let's say it's a more generic one.

>>BRUCE TONKIN: Pick an example and I'll talk it through.

>>BECKY BURR: The morality or the religious names, whatever it is.

>>BRUCE TONKIN: Let's say it was that one.

What they would need to do -- it's not just a case of -- it's not like a public comment. Just saying that word I don't like. They would actually have to provide evidence in the same way, if you were going to object to a trademark on the basis of, you know, being a nasty swear word, right, you actually have to provide evidence to support that objection.

>>BECKY BURR: Okay. So if upper making --

>>BRUCE TONKIN: In other words, it's not the claimant, the applicant that would have to provide evidence in that case. Let's just pick a swear word, think in your mind a swear word; right? I won't give you one. Think in your mind a swear word. And you choose to put that forward as a top-level name, right. That's all you have to do in that particular case. You, as the applicant.

Then someone, to object to that, actually would need to show that that is a commonly-accepted swear word, it's generally accepted as morally objectionable. They'd need to provide evidence. It might be the fact that it's listed in a whole series of laws.

So the work would be on the complainant's side there. It's not -- I just want to make clear, this isn't a public comment exercise where you get people saying, "I don't like that word."

They would go into an external dispute provider and need to be able to prove their dispute before ICANN would take any action.

>>BECKY BURR: Okay.

So on the category that's just the -- you know, it's not the third -- it's not the second, the morality one, but relating to a religion or that kind of stuff.

>>BRUCE TONKIN: Right.

>>BECKY BURR: Would anybody be able to say, "I object to" --

>>BRUCE TONKIN: No, they would not. Taking the religion one, think of a religion in your head.

It would have to be the established institutions, probably plural, depending on the religion, that represent that that are raising the objection, not just, I am a such and such religious observer, and I object.

It would have to be if I am a member or part of a religious community, I'd go and ring the boss of that religious community and say, hey, we should do something about this. And if it's not as centralized as that, you would then say, okay, a number of those religious organizations, maybe they're across a number of countries, would collectively be objecting to that.

>>BECKY BURR: So who decides whether they've met the threshold of being sort of in that category of groups that is entitled to object?

>>BRUCE TONKIN: That's then where it would go -- that's -- that's the hard bit.

But -- I mean, the question is, can we find some external body that would do that for ICANN. Or is ICANN going to need to do that? I don't think we know the answer to that yet.

>>BECKY BURR: Okay. Just one more final question.

In terms of the established morality, who decides what the established morality, whether that objection is valid?

I'm asking the same question, and in each case is, what's the gating? What makes it a real objection that fits into these qualifications, whatever they are, or not?

>>BRUCE TONKIN: Yeah, well, I think you've really got two quite separate questions there.

The question is, who has standing to raise that objection is one question.

And so we're saying in case of religion, it's, you know, people having standing to do that.

In the case of the morality one, the -- it's similar -- the way we've been thinking about it is similar to the way that it's handled in, say, trademark law. You'd have to basically demonstrate that it's a problem.

>>BECKY BURR: So who makes the decision about whether something is a -- contrary to accept legal norms?

>>KURT PRITZ: Yeah, so --

>>BRUCE TONKIN: That's the -- this provider or panel that's doing that.

>>BECKY BURR: Okay.

>>BRUCE TONKIN: And the question is, can we find someone, you know, arm's length of ICANN who can do that and just come back with the answer? I don't know, maybe not. But if we did form that panel.

But the laws -- they'd have to come back with the laws, basically.

>>BECKY BURR: Okay.

>>BRUCE TONKIN: As part of the complaint, you'd have to say, here's the ten laws that give -- that first demonstrate that this is sort of -- this is commonly accepted amongst a number of, you know, national locations, if you like, or jurisdictions. And -- let's say it was a terrorism-related thing, all right, you could probably find a number of terrorist laws that said this is a bad thing, this is a terrorist word and here's all these laws in these countries that are against terrorism. And that's the case you're making to the panel to then consider that.

>>BECKY BURR: In each case, the answer -- my questions are separate, but they actually get to the same thing, which is, there has to be some authoritative body that makes a call. And you're suggesting it's not ICANN.

>>BRUCE TONKIN: As much as possible, I would prefer it not to be ICANN, yes.

>>KURT PRITZ: And one more point is, on that bullet, that language --

>>BECKY BURR: I'm not saying I agree there has to be. I'm just saying if I'm understanding correctly.

>>KURT PRITZ: So that language about "contrary to accepted legal norms relating to morality or public order" was lifted out of an I.P. treaty, the Treaty of Paris in I forget which year, with the -- with not the express understanding, but the hope that there is some precedent and some existing body of decisions around that, so that we'd be relying on something that's already gone before and not creating something new.

>>BECKY BURR: Okay. Thanks.

>>PHILIP SHEPPARD: Perhaps also let me just add to that.

I think part of the discussion that we had in the committee is that that job is not as impossible as it sounds, of course, because, Becky, you're familiar that those sort of decisions are taken every day in trademark offices worldwide, based on sets of guidelines they have there, and I think if you look -- if you look in the detail of the report, there are sort of suggestions that there were -- there are some quotes of some of the guidelines used in some of those decisions in questions of morality which go to the things that are just a question of distaste is not a sufficient test, but something that might cause general public outrage is.

And so that sort of experience and practice exists elsewhere.

And I think all we're saying is we would have a provider who would hopefully have that expertise in-house to be able to do that.

>>BECKY BURR: I appreciate that.

I guess the only thing I would say is that it's not absolutely clear, it's not clear that it's not. But it's not absolutely clear that the same kinds of considerations in a trademark context are the ones that should apply here. And that at least is a question that should be examined.

>>BRUCE TONKIN: And that would be part of that body making that decision, I think. Yeah. So we're not going to, you know, go to a trademark -- trademark office and ask them to decide.

We're more likely to use some sort of international dispute panel that would take into account the applicable laws and the circumstances.

>> YOAV KEREN: My name is Yoav Keren from Domain.net.

I would like to start, pointing out something that needs, I think, a little clarification. It's for you, Kurt, and also Bruce.

You kind of said that the UDRP process is a final paper. And now we have been working in the IDN working group, and there were other working groups dealing for a few months with many issues which I think some of them are not clearly in the same definition as what you have presented generally on the new gTLD. There are some specific points regarding the IDN.

So I just want to make sure that these considerations that we have just filed a week ago will be taken into consideration before there is a final paper of the council.

>>KURT PRITZ: I think Bruce -- I'll let you go on, but I think Bruce was -- it was sort of a euphemism that the final paper has been around for many months, meaning that it's not final.

>> YOAV KEREN: Okay.

>>BRUCE TONKIN: Let me be very specific about it. We spent hours going through that IDN proposal document yesterday. And there were a lot of points coming from that.

That material needs to be incorporated into the new gTLD report in the relevant places. And that's the work that we're going to be working on, you know, through April.

But we do need to converge, because you can actually work on this forever, as we have discovered. And some of the implementation issues will need to be worked on, I guess, further after we've finalized our report.

But what we will be doing is, in the next six weeks, we will be doing a number of edits to the report. And so that final report, I'm saying, is -- we will -- probably the better way of saying is, we will finalize the report in mid-May; right? It's not final now.

>> YOAV KEREN: Okay. To continue -- I'm sorry, I wasn't yesterday. You know that I tried to get here on time for the IDN discussion.

One of the issues that I wanted to raise specifically that I think doesn't work very simple with the general way it's presented here is the notion of confusingly similar. And I know you've spent more than 100 hours on it, so let's just spend another couple of minutes.

The thing I wanted to say here, that the IDN working group, one of the issues that was raised regarding the confusingly similar is that the transliteration or the phonetic similarity between languages should not be considered as confusingly similar, the phonetic part. And this is something that is specific for the IDN. The question is, I think this is something that should be considered very deeply by the council, because there's a difference between languages, and this is something that should be considered.

Now, the other issue is that in the process that you brought up, it's possible for anyone to object, if I understand right.

Now, the problem is, because it's going to be kind of a costly process, we must find a way that it won't happen that the one with the -- that has more cash, more money, will be the one to win.

So if -- if someone applies for a TLD, there comes another body, another company that wants to object to it, they can just, you know, take it forever. And so that's a very important point.

And being more specific on the IDN world, I am -- I think that GAC asked the GNSO to consider that there will be a diversity among the registries, geographical diversity among the new registries.

Well, this comes in also, because if you want to have some registries in other countries with less, you know, economical -- not stability, but sources, economical sources, so that's another point, if someone from a stronger company can object, you won't get that diversity.

>>BRUCE TONKIN: So let me pick a couple of those ones first.

On the confusingly similar, the view is that the issues between scripts other than the Latin script already do exist in the trademark community. And so there is experience in dealing with the issues that you are talking about already today. There's nothing really especially new about different scripts in different languages of domain names that are any different to how they're dealt with in -- outside of the domain name system.

So that's the view there, that there's enough body of experience in trademarks where you have scripts and languages as well.

Obviously, you would use totally different people on panels, depending on their skills. So if it was an issue between, let's say, two Chinese scripts, obviously, the panel would consist of Chinese people, or Chinese-speaking people, at least. So there's an element that the implementation's different, but the principle's the same.

The other comments you make about, I guess, fairness, et cetera, maybe this isn't completely clear -- this isn't an objection to the extent that you then get the string because you objected. So you can't spend money to say, hey, I'm going to come in late, because I want to get -- you're applying for dot hello, and I want to get it as well, I don't get it. If you have a way of objecting to dot hello because it's confusingly similar to something, then no one gets it is the principle there. So it's not really an incentive to go and spend a lot of money going against you.

If it's related to institutions, then, yeah, the idea is you would obviously seek the support of those institutions, whether it was a religious group or a geographical a group of some sort.

The assumption is that you have already invested in getting the support of that group before you put in your application.

>>YOAV KEREN: Okay. I understand.

But --

>>BRUCE TONKIN: And bear in mind again, just to be absolutely clear, let's say you applied for dot bank and a bunch of group of people with money came against you for dot bank, then I can't get T all I can do is stop you from getting it. If they want to get it they have to go into the next round or next process.

>>YOAV KEREN: Maybe something can be proposed here that all or most of the costs of this procedure will be beared by the objector.

So if you object, you should bear the costs.

>>BRUCE TONKIN: And I think that's going to depend very much on what the objection is. So if it's technical it's going to be a bit different.

>>YOAV KEREN: I'm not speaking of technical. I'm speaking of -- All the other issues are not technical.

Thank you.

>>AMADEU ABRIL i ABRIL: I'm Amadeu Abril i Abril, very fond of new TLDs in general.

Two different sort of questions.

First, very shortly, for Kurt, Liz and staff in general. Can you provide a pessimistic time line for the application process?

>>KURT PRITZ: No, I can't.

>>AMADEU ABRIL i ABRIL: Then provide optimistic and abide by it.

>>LIZ WILLIAMS: We can auction the answer, if you like, Amadeu.

>>AMADEU ABRIL i ABRIL: No, the question is that everybody loves good surprises more than bad surprises, and it would be good for lots of prospective applicants, people trying to decide whether this is worth or not, to know something, even the worst-case scenario, or a mid case, bad-case scenario, and also know whether there is some idea about future rounds that would perhaps give off a lot of stress on the next stop in this trip.

The second question is also, can you give a pessimistic estimate of the cost for the application? The answer will be probably not.

I mean, no challenges, no contention, you know. Is it going happy, one of those that you love.

>>KURT PRITZ: Yeah.

I think that an estimate for implementation of a policy would be six months after the policy is approved by the board to populate panels if sufficient work can be done in parallel with the policy development process to develop -- to develop the dispute resolution providers.

So it's sort of dependent on where the policy leaves off and where the implementation picks up and how much work is involved there.

So it's probably not a pessimistic estimate.

>>AMADEU ABRIL i ABRIL: Okay.

>>KURT PRITZ: I think the costs are -- the costs, when everything goes well, are in the five-figure well and the cost when things aren't going well is in the six-figure range.

>>AMADEU ABRIL i ABRIL: And the other question is regarding contention and opposition.

I understand the contention part. Even if I don't agree with all of this, this is where it is.

Now, there is something that's missing, because we are trying to do things that fit for all the cases. And many different applications will be very different.

I mean, and the question here is about how we take into account generic opposition, the massive opposition from, you know, the intended users.

In some cases, I don't care. If I propose dot cyber and you say "i don't like that," that's it, I may fail.

But if, for instance, I propose dot Paris and the Paris local authorities and the French government and lots of Paris commercial, noncommercial and citizen groups don't like me running Paris, it's not the string. It's just me, Amadeu Abril i Abril, running dot Paris, for instance, well, they can go to arbitration. But I can see the French municipality or the French government accepting to go into arbitration for that with a private person or private entity, and especially a foreign one, which is a strange way of solving that.

Let's take for example, the president of dot cat and let's say he come here proposing the names of nonexisting language. Quebecois, which in fact is French. Verlan, which is a funny way of speaking French, inverting the syllables. Inventing language is something that affects public policy, but as I am inventing the language, it is not in the name, you know, any official land of that language anywhere. Nobody has any right to that string.

I am inventing a language, defining the language of the community, for instance. But this is something that is relevant.

So for this case and many others cases I could put into the table we should need to take into account that some sort of organized opposition to either the domain or the registry proposed behind the domain should be taken very seriously in the report of the staff in the evaluation in the first place.

>>CHUCK GOMES: Kurt, if I could respond to that. I know that the working group isn't really finished with this particular area, but considerable work with regard to who has standing on this, a lot of thought has been put into that. It's not something that the working -- that the committee has ignored. And I think there will be some -- some of that work is going to flow into the full committee in the coming weeks that will be pretty consistent with what I'm hearing here.

>>S. SUBBIAH: This is Subbiah, idns.net.

I understand that Bruce mentioned earlier that all the IDN issues have not been put in this report yet, this report, the new gTLD report. But one of the things I wanted to highlight is that there was a discussion in the IDN working group and there was some support to the notion that, in many of these things, that the inclusion of the language communities shall be -- local language communities should be brought into the picture.

So one of the suggestions that was made, and I'm not sure how it's going to fit into this policy changes, is from the perspective of it's not just -- there are two parts to the language community input. One part is in the sense that during the time of evaluation, you know, some feedback is taken from the language community, I think there's some agreement to that.

Now, -- working with the language community.

But I suspect for legitimacy reasons, just in general, acceptance reasons around the world, that it would make a lot of sense from the selection committee point of view, there may be standing committee or whatever, but that in each language case in IDN gTLDs this is general policy for everything. If a particular language is being picked as a candidate for a gTLD string, then as part of the selection committee itself, there should be inclusion of individuals, at least some individuals, from the regions. If there's a diaspora as well as a country where it is mainly spoken, at least one person from both or something like that. That would conflict with my second point which is from the way that this is being presented in the way of cost, I heard figures like $100,000 being thrown around for even just a not-too-difficult application.

But if you were to convene people from the communities, then it would necessarily imply that the cost may even go higher. And once that happens you are basically preventing people from the language community to participate for financial reasons.

And I think, really, many ccTLDs today, ASCII ccTLDs today that run large numbers, hundreds of thousands of names, operate on a reasonably technical sound level around the world for less than 40, $50,000 a year. Many of them do. We all know that. And these are the countries from which a lot of the new IDN TLD applicants will be coming from.

In the context of taking only that much to operate an existing very stable ccTLD to then say it costs a hundred thousand or more to apply, it's not a way of achieving anything that the previous speaker mentioned, that GAC would bring up in terms of diversity of applications from the regions.

So I understand that ICANN has to have a fair policy for everybody in all these things, but certainly a question needs to be addressed. So two parts of this that I want to bring up.

>>KURT PRITZ: I think your comment about language participation on the panel is very fair.

With regard to cost, we're trying to develop a process where applications that weren't subject to objection or controversial or met the criteria would be accepted in the cheapest, fastest way. And that applications that were more problematic were identified early, and the costs were identified early so decisions could be made.

But just as it's meant to identify objections, it's also meant to help applications that aren't controversial or problematic in any way and be as inexpensive as possible.

>>S. SUBBIAH: Just a point to that --

>>BRUCE TONKIN: Let me add one other comment to that, Subbiah, because this has been raised. I guess cost obviously is an issue if you are considering putting an application in. A number of people have asked the cost question in a number of different ways. But the other thing we have been thinking about is the cost of the first round, Kurt's slide is saying the startup cost of setting all the process up is expensive. And it's not dissimilar to, when, at the second level of new gTLDs, they have often had a higher cost often in the first few months for applications, because they are dealing with all the legal issues, et cetera, et cetera. And then after that initial phase, most new gTLDs have dropped their prices substantially. And we've seen that with pretty much every new gTLD that's launched has had higher cost in the first few months compared to a year later.

And what I mentioned is ICANN is trying do this on a cost recovery basis, let's call it.

You can imagine that in the first round, the costs would be higher because they are setting up all the process and I think as Amadeu said, as long as we announced the first round was January of '07 and the second round was, I don't know, first of July '08, then, one, you are telegraphing the fact that at a particular date there's going to be a second round. And secondly, you might also telegraph that the costs might be lower in the second round because by that time, a lot of the costs are sunk in investment. You have already spent the legal work for doing the RFP, so hopefully the first round pays for that. So then when you get to the second round, you are just talking about the marginal or incremental costs. And as Kurt said as long as you have ticked all the boxes and done it right, there's no reason why that should be a high cost.

>>S. SUBBIAH: May I propose a radical solution?

ICANN has been going for a bunch of years now, eight, nine years longer.

Now, it would seem that the whole principle here is cost recovery.

Why does it necessarily have to function that way? Before -- I want to add a couple of points. That is when ICANN was a younger organization and starting up and were on the whole a few million dollars, that was perhaps a reasonable thing to do. Now there's a lot of revenue coming from existing registries, ICANN's budget has expanded quite a lot, so it could be possible that cost recovery is not the main mechanism, meaning you charge a nominal fee, because you don't want somebody to apply is to make a frivolous application. Objection is a different issue. So you could keep it something reasonable, 10,000, 20,000, something reasonable, fair to everybody, not just the people coming from poorer countries. Fair to everybody, something that's a bar everybody can meet.

Now, of course you collected some, but then there's a shortfall for ICANN if you are thinking in terms of cost recovery. Well, that can be recovered either from, you know, the registries themselves going out further, the one that wins and makes some money, you adjust the price a little bit. I think most people in the world can afford a few dollars for a domain name. I think the price of a domain name is not the real issue these days.

So that could be recovered.

Then there's a completely separate issue. If you are talking about in terms of a startup. These companies, see, what you're saying is companies are coming to apply for new gTLDs. You take the risk, you put up the cash and you be a startup, right? And you put up the money and if you win, you recover yourself.

Well, the same question could be asked of ICANN, because ICANN has been around for so many years and it has collected a lot of money in the first rounds of all these registries, so ICANN ca put up some of this money towards helping the next phase. It's sort of a sharing business here where the applicants are still required to put up some amount of money. It would seem that would be a reasonably fair way to go. This would avoid this whole question of relative cost to different people. Just a point.

>>BRUCE TONKIN: Just one comment on the relative cost because the noncommercial users constituency has made some missions in this area and we actually have in our guidelines at the moment that there should be some mechanism to assist organizations where cost is an issue. And let's say it's a language community in the area of the world where the economy is not strong.

One of the other ideas that has come up is maybe the surplus from the first round is then used to establish a fund of some sort, could be a trust fund, and that fund is used for those specific purposes. So it's not as if we haven't thought about that idea and we are certainly looking at it, there is already text in the report and I encourage you to look at it.

And maybe work with a noncommercial constituency that submitted the original text. Because we're certainly trying to put something in the report about -- of that nature. Not so much a policy issue, but it's more sort of an implementation guideline for ICANN to think about.

And again, I want to stress that this is an ongoing process. It's not just a once off.

And so trying to keep the first -- trying to create granting keeps and things like that in the first round when we haven't even got the rest of the things worked out yet I think is high overhead.

But on an ongoing basis, there's certainly an intent to be able to support the notion -- I guess the fact that in order to encourage diversity in things, just like a number of other activities around the world, you need to create mechanisms for supporting that.

>>S. SUBBIAH: Thank you.

>>BERTRAND DE LA CHAPELLE: Hi, I'm Bertrand De La Chapelle from France.

First of all, I want to say how much I appreciate the presentation that was made, especially on the historical background and the evolution of the discussion.

In particular, because it explicitly mentions the fact that it's a public resource and the importance of the commitment and support of a given community to the actors that will manage a given string or a given domain.

But I wanted to make two comments regarding two alternatives that Bruce has presented as if they were either/or.

In recalling or making a recap on the history of the discussion regarding pro or against new TLDs, there was basically a set of arguments in favor of the creation of new TLDs and a set of arguments against it.

It can be presented as an either/or. The reality is that as a public interest objective, there's a combination of both. You want to enhance the positive aspects and minimize the negative aspects, which is actually providing an interesting guideline in all the evaluation processes later on, according to the criteria you described yourself. Does it increase choice, diversity, competition, provide a space for communities, and at the same time is the string that is introduced minimizing the obligation of defensive registrations and so on.

The same reasoning --

>>BRUCE TONKIN: Before you go further, I just confirm that that is exactly how we have been looking at it.

Bear in mind the presentation is a simplification of eight years of discussion.

>>BERTRAND DE LA CHAPELLE: I could imagine, and this is why I was raising that point. Because it applies equally to the second alternative that you have been presenting, that basically says either you get a structured taxonomy, and I hear a word that was not mention but that what seems to be implicit which is an a priori taxonomy, or you have -- you allow communities to self select. And here the implicit is made explicit, which is market-based approach.

The two elements that are implicit or explicit are not a direct logical consequence of the main expression.

You can have a structured taxonomy that is not a priori, and you can have a community that self-selects through other mechanisms that market-based approach.

Which means that applying the same reasoning as the one we had to see whether there should be TLDs but, yes, in what proportion, you can here consider there would be a path or an objective is that the community, as a whole, and the communities find ways through this process to progressively elaborate an evolving taxonomy through the combination of emerging proposals and a regular thinking, like you did in the presentation, whether a given string is actually representative of a type of string.

I think this is exactly what we're all trying to do. And once again, the presentation made it even clearer, I think, in the discussion.

The most important thing in the presentation was the description of the goal, the "why" slide. And I would like to ask you how much of this is formally agreed formulation within the group, or how much was a presentation here?

>>BRUCE TONKIN: Let me answer that one specifically.

In terms of what we've attempted to do in the report at the moment was collect the diversity of reasons for. So at the moment, the approach we took -- I mean, I am quite happy to take that back to the group and see if they want to make that the one single reason for, but the way I was trying to present that is that is a reason, and what we have in the report is the -- a number of reasons that were put forward.

So the report just basically says here are some of the reasons that were put forward. And some of them weren't commercial, and we included that. We are trying to recognize there is a diversity of views.

>>BERTRAND DE LA CHAPELLE: As a matter of fact, this is exactly the point because from what I understand in the discussion and that quote, you now have at least two positive reasons for the introduction of gTLDs, and therefore, two positive criteria to use in an evaluation process.

Without specific order, the first one is the traditional one regarding competition, diversity and so on. And the second one was very nicely formulated to allow globally distributed communities the opportunity to identify themselves at the top level.

In that respect, I am referring to your answer to Becky earlier. I do not see exactly where the swear word might fit in one of the two categories, but that's an issue. Creating a string with a swear word doesn't serve either the community of actors or competition, per se.

>>BRUCE TONKIN: Okay. But what you consider to be a swear word might be different to what the applicant considers to be a swear word; okay?

>>BERTRAND DE LA CHAPELLE: Okay. That's fair.

>>BRUCE TONKIN: So I am not proposing the registration of swear words. I'm just saying that a word that might actually represent a community somewhere in the world might also be considered to be a swear word by others.

And therefore, that's the --

>>BERTRAND DE LA CHAPELLE: Okay. Just to finish because I am already speaking too long. The point I wanted to make is that the process for evaluation, and particularly the one that is described here on the slides, is, according to what we just discussed, missing one component which is the positive evaluation criteria.

There is an evaluation on the technical and economic dimension. There's an evaluation on the non-objection, the negative criteria.

I think the whole discussion shows that there are a certain number of positive criterias regarding the benefit for a given community or the community as a whole of introducing a given string. And in that respect, the evaluation, by whatever panel, has actually three different subsets. One is the opportunity to introduce the string itself, second one is the choice of the applicant, and the third is the operational methodology.

I just would like to stress how important it will be in the future discussions to discuss what are the criteria and how the panels are going to be composed in the future.

>>BRUCE TONKIN: Just one further comment on that first item there about the criteria.

One of the things we discussed on the weekend was actually identifying what we think the measures of success would be.

Because one of the other things we have been thinking about is how do we know in ten years' time whether we did a good or bad job. What actually were our measures of success?

And I think dealing -- the way I presented the for and against, if you like, what would actually be a measure, let's call it a failure, is I create these TLDs and all they have in them is defensive registrations. I would say that's expressed negatively as a measure of failure.

A measure of success, just picking that particular reason I selected, but a measure of success is a genuinely new cultural group, such as dot cat as an example, and it's being used in that community. You know, that community is actively using the name would be a measure of success.

So one way I suppose of responding to your suggestion that we have been thinking about ourselves is to articulate what we think the measures of success would be, and also that we actually have a review process built in that we review in six or 12 months' time, or whatever the time frame should be, and see whether we are meeting our objectives.

>>MILTON MUELLER: I love you, Bruce. The answer to bureaucracy is more bureaucracy.

[ Laughter ]

>>MILTON MUELLER: It's just wonderful to watch this process gathering steam.

But I have a lot of comments about this process, as you might know. But I want to start with a very simple question. Just it's a very simple question, and I want to see how you -- not just you, Bruce, but anybody who participated in the sewing together of this Frankenstein monster would answer it.

And the question is, is it possible for anybody to create a controversial proposal and get it approved under this process?

>>KURT PRITZ: Yes. So the answer is yes. I think that it's probably a long and costly process. I think it has to be defined for the applicant up front. But also, that we need to -- There's lots of noncontroversial strings and applications out there.

>>MILTON MUELLER: Let me just understand first how it would happen.

So if I propose dot abortion and the Catholic church weighs in against me, do you tell the Catholic church to take a flying leap or what?

>>BRUCE TONKIN: Well, you work it through the process.

And what's the basis for objecting to that? Is it on the basis -- Is there a law against abortion? There is in some places. I don't think it would necessarily be the Catholic church that would be able to do it on the basis of being a religious organization.

>>MILTON MUELLER: Anything that is controversial, there is going to be a law against it somewhere. Is that not right?

>>BRUCE TONKIN: Well, there's two elements to it.

Maybe I'll answer your first question because I don't want to go down -- I can see myself getting into trouble on that particular topic.

But -- or the particular example, I should say.

But let me put it another way.

Controversial strings have already got into the root. Every TLD that -- in fact, many of the TLDs in the second round had significant objection, and some of those have got through, so I think the answer to -- I'm using controversial in the sense that there are objections to it, yes, it can get through. If you are going to choose particular words, I don't know the outcome because -- but as a category, is it possible for something controversial to get through? Yes. In fact, it already has. And I'm not saying --

>>MILTON MUELLER: Wait, wait. They weren't using this proposal. They were not using this proposal. So that doesn't really answer the question.

I'm asking under this proposal, which seems to be inviting anybody in the world to object to something that they think is against their values somehow, their legal --

>>BRUCE TONKIN: But there's a fairly -- there's a bar to do that, though. That's the difference.

>>MILTON MUELLER: So what is the bar?

>>BRUCE TONKIN: Well, the bar, taking the example of -- let's take the bank example. What I'm saying is the objector to that -- let's say someone applies for dot bank. The objector is that is not just a bank objects. They would have to show they have the support of the banking community to object. So in other words, we're flipping it around to a degree. We're not requiring them --

>>MILTON MUELLER: In that case, you're talking not so much about semantic controversiality, you are talking about a proper matchup of the applicant and the idea represented, which at least has something to say for it.

But it seems to me that you have opened the door wide to purely semantic or sensorial types of objections here, with your discussions of morality and public order and your invitation for so-called legitimate groups to express objection basically on just value conflicts.

>>BRUCE TONKIN: But what we are saying there is there have to be widely accepted values, if you like, internationally.

So it's not just a --

>>MILTON MUELLER: So it's the clash of civilizations, is it?

>>BRUCE TONKIN: It could be, but the particular one that you -- the particular example you have suggested, abortion is probably something that you'd -- that would be something that would need to go through the process and we would see what comes out of it.

But the question you are asking is is it an internationally accepted legal norm not to have abortion. I don't actually know the answer to that question. But if it was, then yes, I think that probably is basis for objection on that particular word.

>>MILTON MUELLER: And I think that's tragic, that you are basically saying -- you are creating a political process of censorship.

You're sort of abandoning 300 years of liberal ideology about freedom of expression and saying that we are going to decide what is allowed to be uttered at the top level based on an alleged universality that doesn't exist. And I would just remind you that one of the ways that we ended several centuries of religious warfare was not by deciding which religion was right; it was by the principle of tolerance, which allowed all the religions to exist and separated state power from expression and conscious and belief. And that's, I'd suggest, a direction we have to go. I can go and registering FUCK. See if it goes up there. In the second level in a number of domains.

>>BRUCE TONKIN: Yes.

>>MILTON MUELLER: And the world didn't fall apart. And I can register Jihad, and I can register -- I would invite you all to look at what's under God.com or Jesus.com. We didn't have massive global political fights about who got those domains. There's not that much difference. There are some differences, but there's not that much differences between the top level and the second level unless you make them into issues. You can create a political conflict and a need for political conflict if you try. I mean --

>>BRUCE TONKIN: Let me be clear.

If you insert that word that you've just suggested, I'm certain you'll have political conflict.

>>MILTON MUELLER: At the top level. And I'm suggesting.

>>BRUCE TONKIN: At ICANN. You absolutely will. I think there are a number of governments that would certainly find that -- yeah, there would be political conflict. I'll leave it at that.

>>MILTON MUELLER: Well, my mother would not approve. But that's neither here nor there.

The point is, you are a technical coordinator of the top-level domain name system. You're not an arbiter of global speech norms.

>>BRUCE TONKIN: Yeah. The other way I'd answer that, Milton, and this is just speaking personally, I think the difference at the moment, as opposed to where it might end up, but the difference at the moment is, there's a relatively few names at the top level. And if I just add one single name, I don't add any other names, and I add the name you've suggested, I think that would be controversial.

If you add that name at the second level to dot com, it is one name amongst 40 million, and that particular word is probably spelled already in a thousand different ways already. So adding another way is a sort of drop in the ocean.

So I think that would be why I think there would be political issues right now, because we are creating -- we have artificial scarcity there, and you add the word at that level, it kind of stands out. That's just a personal view.

>>MILTON MUELLER: Just one other factual point. That is, when you look at the proposed grounds for an objection to a string, all but the second one there has perhaps something to say for it. And you've been invoking this international treaty, which is basically a trademark treaty. And the thing about trademark is that it's not a standard for determining whether you get to utter a word or whether it gets to be published. It is a grounds for determining exclusivity between two applicants, so that if one applicant's use of the term encroaches on another's, then you have some basis for saying, you get it and you don't.

But you do not have that as a basis -- there's no legal grounds, there are no universally accepted legal norms for saying you can't have that word at all. Nobody can have it.

And so you cannot invoke this existing body of international law to justify what you're doing here.

>>KRISTINA ROSETTE: Bruce, if I could respond briefly to Milton to. A certain extent, that's not necessarily correct. And to the extent that we were looking at how the trademark offices were operating, we were looking at more let's look and see how organizations, how governmental organizations have tried to adopt standards that can be applied universally.

In the particular context of a mark, for example, that would be refused registration under the grounds that it's immoral or scandalous, no one is going to get that mark. You may not; I won't; Liz wouldn't. No one will. Because the determination will be made that it is in violation of whatever standards the trademark office has applied.

So that's really kind of the basis that we were coming from when we tried to refer back to that language.

And we're certainly amenable to -- if there are other standards that have been used in this context, we would certainly, I think, be welcome to considering those.

>>JORDYN BUCHANAN: Hello. I'm Jordyn Buchanan, and I'm representing only myself.

First, a quick observation in regards to the point you just made, Bruce, which is, it seems like you're saying that because there's scarcity -- scarcity causes a bunch of problems and makes us think really hard about all of these issues. And you're suggesting that the solution to that is actually to think a lot harder about these issues and continue to be involved with that very difficult process.

It seems to me the obvious solution is to get rid of scarcity and then you don't have any of the issues anymore. So a much easier and more open -- the most. Sorry.

The most open approach possible to making it possible to add a large number of new TLDs relatively quickly would do a much better job at resolving these issues than anything we're going to do with a complicated procedure that we're going to have to funnel everyone through. A comment.

Second, a question. Has anyone looked with regards to this point about trademarks and property, legal rights of other parties, has anyone given any thought or done a survey to figure out how many strings that's likely to affect?

Like, how many trademarks are we talking about here that we're going to be protecting?

How many trademarks -- how many strings are potentially covered by existing trademarks?

I think that's -- seriously, how many trademarks are recognized in international level that would be covered by this rule?

>>UTE DECKER: If I may answer that point very quickly, I think there's been a lot of concern that every possible string, if I understand your concern correctly, every possible string would be challenged by a trademark holder, because what combinations of letters are there that are not already trademarked in one country or another.

>>JORDYN BUCHANAN: Really, I guess, my question is have you looked to see if there's a universe of useful strings that are not covered by some existing trademark somewhere.

>>UTE DECKER: Not being a trademark attorney myself, but a member of the IPC, it is my understanding that generic terms cannot be protected by trademarks. So, say, for example, the combination of TRAVEL could not be protected by a trademark, because it's a generic term. And in that sense, I think that's almost a -- that's almost a positive relationship in the sense that anything that would be interesting for any applicant for possible gTLD would be, more likely than not, not trademarked, because it would be generic.

>>JORDYN BUCHANAN: I think that's true when it's applied to the thing that it's describing. But -- so I can't create a brand of apples and call them "apple." But I can certainly create a brand of computers and call it "Apple." And so no one is allowed, then, to register the string apple because someone decided to call their computers that.

And similarly, I couldn't make -- I couldn't make a brand of electronic little bugs called beetle, but I could certainly call a car a beetle. And people do that as well.

I guess I'm just curious as to whether anyone has taken a look to see whether or not -- what the proportion of strings in the universe that might be -- that people might consider to be useful are versus the strings in the universe that already trademarked.

>>KURT PRITZ: I don't think the registration process prescribes the registration of a tld dot Apple for the apple growers association.

The process is meant to prevent somebody going to an island country and registering names or having local law changed to help aid in the -- or facilitate the registration of certain names.

So that's why there's a distinction here about international law.

>>CHUCK GOMES: Jordyn, can I respond to you in a little different way?

>>JORDYN BUCHANAN: Sure.

>>CHUCK GOMES: We heard your question. And it's a good question.

But, first of all, I want to point out that the protecting rights of others working group is not finished and won't be until May. So right now, we don't know what direction that's going to go. Okay? It's an issue that's been raised by members of the community different than some that have been at the mike, okay. So it's one we need to consider.

Secondly, it's important to point out that in -- with regard to the process, one of the things we've tried to do is to flag possible concerns very early in the process so that applicants don't get drug out over two years and have to go through a trademark dispute or some other rights dispute and so forth. So there are some other values besides just protecting the rights of particular I.P. holders or some other thing.

And, again, I think it's important to keep those two things in mind. We're not finished yet. And, secondly, there are some other values to dealing with these things up-front rather than going through an expensive, drawn-out process.

>>JORDYN BUCHANAN: Sure, okay. I'll make two very quick points and sit down.

First is, I think the language that's used right now says registration or use of the string. And I think that's distinct, Kurt, from what you're talking about. What you're trying to prevent is a situation in which the domain is used in a way that might infringe. And I think that's distinct from just anytime you register the string, if there's some overlap with a mark, then I think that's more problematic than I think if you're concerned about how the string is going to be used.

The second is I'm just asking a factual question. Before this becomes the criteria, it seems like we should make sure that it's not actually going to exclude, essentially, every useful string that anyone would possibly -- so I just think a little data would be useful on this point before it becomes part of the policy.

It may be that instead of saying this is a grounds for objection that a necessary part of the process is that we evaluate the, you know, proposed string versus the inevitable trademark conflict that's going to happen and we need some way of deciding whether or not we're going to proceed, even though that conflict is likely to exist in essentially every case.

>>BRUCE TONKIN: So let me -- firstly, they have to be able to prove that in the dispute resolution. And I think one of the things we have been talking about in that dispute resolution was how that's crafted.

The -- please bear in mind that those are just sort of summary slides and there's a lot more in the actual report.

But just on your generic word thing, there's a difference between a word mark and an image. So quite often when you see something like orange, that's not a word mark, it's actually the picture that they've protected. Because they wouldn't have been able to get protection for the word.

>>JORDYN BUCHANAN: I think you'll find that's not the case and that both are actually trademarked in many cases.

>>BRUCE TONKIN: Depending on the country, yeah.

But I think the other element of that was, there are some other words that come out of the UDRP, and it's something -- and I've forgotten the exact words. But it would be more than just, hey, there's a string and it's got a trademark and you get it.

It would be the combination of things.

And so I think the intent, as -- in terms of our design of the process, as opposed to how we implement it, but it's not the intent to be protecting generic words, the intent is to stop something like dot Microsoft or, you know, dot NEC or something like that, and then where the use is clearly intended to trade off someone's brand. And it's not intended to be able to create an environment where, yeah, you go and get a word mark in a country that doesn't have strict rules and use that as a basis to block the string.

>>JORDYN BUCHANAN: I just think it's -- ultrasound probably find the NEC, letters NEC in some language actually has some useful meaning and now you're excluding that use as well.

>>BRUCE TONKIN: Possibly, yeah.

>>KURT PRITZ: Peter, are you the last one in line, or is there anybody else?

>>PETER DENGATE THRUSH: I just have a procedural question. This session was supposed to finish 15 minutes ago. I'm on the panel and there's a group of us who are supposed to be on the panel. How long are you going to go on for? Can we go to the bar and when should we come back?

>>BRUCE TONKIN: We're actually going to ask if you were the last speaker, I think.

>>VITTORIO BERTOLA: Okay. So I'll try to be very quick. I'm Vittorio Bertola --

>>PETER DENGATE THRUSH: How long are you going to go for?

>>BRUCE TONKIN: We want to finish now.

>>KURT PRITZ: Let's say three minutes per comment and there's three more comments.

>>PETER DENGATE THRUSH: Another quarter of an hour.

>>BRUCE TONKIN: I think we should stop at this point.

>>LIZ WILLIAMS: I'm happy to take the comments and you can put them here.

>>BRUCE TONKIN: Vittorio's the last speaker, and we're -- can't keep on running over other people's sessions.

>>VITTORIO BERTOLA: I just wanted to point out a couple of things. First is that, having been involved with, I mean, a specific application that's currently going on, it's impossible to determine an international set of values.

And it's -- I think it's also impossible to go by either end of the spectrum. So either take the intersection of what's acceptable to everyone or the union of what's acceptable to everyone. It's always going to be a call. And it's always going to be a subjective call when something is controversial.

So the thing I would like to recommend is to keep every policy decision like this into a policy body. So let's not pretend that it's going to be a technical evaluation, it's going to be staff or it's going to be independent panels. It's always going to be in the end a policy decision whether to approve or not to approve something, especially when it's controversial. So it should stay, I'd say, with the ICANN board or with, anyway, an entity, a body that has representative -- is representing the different parts of the community.

And the other thing is about cost. But everyone's already said that. I just wanted to point out that there are people who have been giving away and running registries -- giving away domain names for free for ten years now, and it's working fine.

So the -- the only barrier towards these people getting a TLD is that the cost you might want to put into these applications. So, actually, you should keep that as low as possible to facilitate everyone. Thank you.

>>KURT PRITZ: There were two more questioners. And if you could limit your questions to just a very brief period of time, I'd appreciate it.

>> KARL MANHEIM. Karl Manheim, Loyola Law School, Los Angeles, currently at the University of Bologna.

I was cited in your report as one of the authors suggesting that you move towards auctions for the allocation of top-level domains.

I didn't quite understand why the auction process was relegated to kind of a last resort when what proceeds it seems to be the beauty contest type of evaluation proceeding that the Federal Communications Commission abandoned after 70 years of proven nonsuccess.

It seems to me that you're recapitulating that experience with the FCC and that an auction process might not only save the $100,000 per comparative hearing that you were estimating, but also move this process along much further.

That's all I wanted to say about that.

I wanted to say one other thing, and that is, as a lawyer and a law professor, I couldn't help but being impressed by how juridical this whole process seems to be becoming. Obviously, ICANN evolved a long time ago from being simply an engineering standards body to setting basic public policy. But now it seems that it's going to be establishing judicial policy as well. And what I've heard described here today is nothing less than the creation of an entire judicial system to determine international legal rights, norms of morality and public order. These are the types of things you would ordinarily expect courts to do.

And with the comparison to trademark disputes, of course, when trademark bodies make these decisions, they're usually subject to judicial review. And I didn't hear that in this proposal, that the -- the decisions from the -- whatever this independent review body would be, whether they would be subject to judicial review or not.

And they're not, that essentially makes ICANN a judicial body with the final say.

And I just -- you know, without -- there's no time, really, to talk about this. But I just think it's important for ICANN and the GNSO to keep that in mind as they move forward, that they're taking on a much more judicial character, as well as the public policy character they have in the past.

Thank you.

>> TAN TIN WEE: My name is Tan Tin Wee, and I come from the National University of Singapore.

And I just wanted to get some assurance from ICANN whether there will be -- with regard to the technical evaluation process, whether groups that have significant problems achieving these high standards of registries and registrars will have some consideration that will take into account that they may not need that high level of and that high stringency of technical qualifications in order to get a TLD in IDN.

The reason is very simple, and some of you may remember why is it that in 1998 that I decided to set up this whole thing about IDN in motion back in 1998. And that was because of the serious concern of the digital divide. In many countries, there are children who would like to benefit from the Internet, and they can't because they're locked in because of this language divide. And the accessibility problems or barriers that prevent these guys from benefiting from the information age is fairly significant. And it was because of that that I set out to start this IDN thing.

So I'm not very happy that it has taken eight years to get to here but happy that we have reached this stage, that we are just on the verge of possibly looking at issuing IDN TLDs.

To give one example, I have some students who are actually volunteering -- I teach in a university. So I have some students who actually volunteer in the Peace Corps type of things to work with Cambodian children; right?

English not so good. Problematic. They're trying to help them along in primary and secondary school education.

So for these guys, our students now have a problem because they cannot apply, because they don't have the resources to look into how to deploy Khmer TLDs for these guys to be able to surf the Web in their own language or to send e-mails to each other in their own language. I'm sure Norbert is somewhere here in the audience, and -- oh, yeah, sorry. Panel. I am wearing my reading glasses.

And suppose he weren't there to push for this. Then we have a serious problem, because then these high-level software standards in technical evaluations and because the financial requirements will now become the other barrier. First we have the language barrier. Now we solved that, technically. Now we have the financial. And then, hopefully not, the technical barriers.

So I sincerely request the august committee members here, the council members here, to seriously take into consideration this issue and seriously address this problem, are we creating another -- pushing another new barrier for these -- this disenfranchised group. And I certainly hope and trust that this community here and all of you here will actually take this into serious consideration and have some assurance going forward.

Thank you.

>>KURT PRITZ: Thank you. Thank you, everyone, for your participation. Thank you, Bruce.

>>BRUCE TONKIN: Thank you. That's the close of this session. And the next session, I believe, is on the topic of the GNSO review, which I assume will probably start in about five minutes, at half past 4:00, to give enough time for the panel to swap over.

[ 4:25 p.m. ]

© Internet Corporation for Assigned Names and Numbers