Olivier Crépin-Leblond: Thank you very much. Good morning, good afternoon and good evening, everyone. This is a call on Tuesday, the 15th of February, 2011. Time is 13:08 UTC, and it is the Community Call on the Interim Report on the Internationalized Registration Data Working Group.
With us today we’re very lucky to have Steve Sheng, who is the ICANN Senior Technical Analyst within the Policy Department, and who will talk to us about the currently open public consultation on the Interim Report of the Internationalized Registration Data Working Group. And the report was published, I think there was a comment period open on the 15th of November which was meant to close on the 13th of January, 2011. It has been extended to the 14th of March. So we are well within the time to provide some feedback.
But first let’s start with a quick roll call, and then I’ll pass the hand over to Steve Sheng. Roll call, please. Gisella? Okay, we don’t have Gisella.
Cheryl Langdon-Orr: She must be muted or we may need to dial out.
Olivier Crépin-Leblond: She might be muted, yep.
Heidi Ullrich: Hi, this is Heidi. I’ll attempt to do a roll call. We have with us Olivier Crépin-Leblond, we have Cheryl Langdon-Orr. We have Cintra Sooknanan, we have Dave Kissoondoyal, Dev Anand Teelucksingh, Sebastien Bachollet, Marc Rodenberg; and from staff we have Gisella Gruber-White when she is around and we have Matthias Langenegger and myself, Heidi Ullrich; and of course Steve Sheng, the presenter. Have I missed anyone?
Okay, thank you, Olivier.
Olivier Crépin-Leblond: Thank you very much, Heidi. And now I’ll pass the hand or the baton over to Steve Sheng, who will be giving us a presentation. Thank you, Steve.
Steve Sheng: Thank you, Olivier, and good morning, good afternoon and good evening, everyone. I am the staff support for the Policy Department for the Internationalized Registration Data Working Group, and we would love to hear the At-Large community’s feedback on the Interim Report.
So the second slide, the goal of the presentation is to first brief the community, the At-Large community of the Working Group’s current work. And the second is we have a public comment period open, and we would very much appreciate the feedback from the community under proposed recommendations.
So just kind of the terminology: the IRD. The registration data are the data that the registrant permits, that include basically the compact information and the development information for when a registrar registers a domain; and there is currently, the internationalization of that has been studied. So that’s our terminology.
So why are we talking about the IRD Working Group? Currently today there’s no standard that exists for the submission and display of the domain registration data. There are guidelines for the IDNs, the domain name part, but there are no guidelines for the rest of the data. So the Working Group was chartered to study the feasibility and suitability of introducing the submission and display specifications to deal with the data.
So there are two questions here – there’s one, the feasibility; the other, the suitability. So the question, the suitability part is that traditionally all this information is in US-ASCII, because since the early ages the English language has been spoken predominantly, those speaking the language with US-ASCII the script; but as the internet has become more and more internationalized that is no longer a safe assumption.
So a question the Working Group tackled is “Is it desirable to have it in multiple languages and scripts?” And the accompanying policy question that comes with that is “If we have it in multiple languages and scripts, how do we avoid the tower of babble effect, where you could have the domain registration data in ten or even a hundred languages or scripts, but people may not be able to communicate with each other when trying to resolve domain name issues. The second part is the feasibility issue – whether it’s technically feasible. So those are the two issues.
As I mentioned before, the domain registration data are data that registrants provide at the time of registering a domain, and the information is often referred to as the WHOIS data.
So the Working Group discussed ways of internationalizing those data, and they adopted a two-prong approach. The first prong is that for some of these data there is already an existing standard on how it should be internationalized. So for example, the domain names, as I mentioned earlier, they’re already standard. So the standard label, the standard Unicode label for the IDN name and the A-label – there’s corresponding information of that label that’s used in the DNS.
So it’s the same for name server names. Sponsoring registrar, this was one field we thought could be remaining in US-ASCII because there is a need for law enforcement, and for example, intellectual property owners, to contact around the world. And the discussion there is to have it remain US-ASCII. Telephone number, fax, then there’s the ITC standard, UPC E.123 to internationalize those. Email address – the email will likely need to become more internationalized to actually contain the U-label, so there’s an existing RFC for that, the same with registration data and dates.
The heart of the issue that we’re talking about and that the Working Group really would like to seek feedback on is how to internationalize the registration contact data. So these data include the admin contact, the technical contact and the domain name owner contact; and this includes names, addresses.
The Working Group came up with a few models. Before going through the details of these models I want to highlight the principles that are gathered inside it. The first principle is to avoid how badly this affects the registration data. We want to accommodate scenarios where a user may wish to enter this information in his or her own language, but we also want to avoid the tower of babble effect.
The second guiding balancing principle is one to balance between the cost and usability of the registration data. So we could consider a range of submission scenarios. For example, a translation would probably be more accurate and more usable as well, but it also bears what could be a significant private cost to the registries and registrars, and potentially to registrants. I use the term “private costs” because it’s a cost bared to the parties compared with a public cost.
And the third guiding principle is we want to consider the needs of human users as well as legitimate automation. The WHOIS information before was primarily for human users, but these days there are lots of applications that are parsing and trying to halve parts and analyze registrars.
So what are the four models? The first model is registrants provide the domain contact data in the “Must Be Present” script. So this is more kind of today’s standard core model, where in some IDN registrations, registrants for one page input his or her own contact information in the local language. And then on the next registration page he or she has to provide that in US-ASCII. So this is one Russian example and one Chinese example in the “Must Be Present” script. To get the “Must Be Present” script, maybe they could do translation or transliteration; that’s really dependent on the policy of the registrar.
The second model we considered is registrants provide the data just in any registrar-accepted script, and it’s the registrar’s responsibility to provide a point of contact to deal with converting these, either the translation or the transliteration request. So for example, in this Russian example we have a registrar point of contact with the URL Nic.ru with phone and fax and email numbers, but the actual registrant information is all in Russian. So that’s model two.
Model three is registrants can provide data in any registrar-accepted script, but it’s up to the registrant’s responsibility to provide transliteration tools and publish in the “Must Be Present” script. So this means then that the contact information in model two which was in Russian has been transliterated into US-ASCII. So that’s model three.
And model four is the information has been translated instead of being transliterated, so I want to call out the highlights of translated versus transliterated. I want to just put two examples hand to hand. I think model one can actually provide a good comparison. So for example, the Russian example, the country name is “Russia” in translation, but in transliteration it’s “Rossiya.” The address is “Vladimir region” in translation, but it’s actually Vladimirskaya oblast. It’s the phonetic way of how it’s pronounced, so there’s lots of inaccuracy there, but nevertheless the benefit of it is that this could be less costly.
So let’s just summarize the models and think of a 2 x 2 table. On the column is whether we do transliteration or translation; on the row is whether it’s the registrant’s responsibility or it’s the registrar’s responsibility. So we summarized model one, three, and four in this 2 x 2 table. Model two of course is not in there because registrars provide a point of contact.
So these models are the primary parts where the Working Group especially would like to seek feedback from the community. We’ve discussed about these models but we haven’t really decided which of these models the IRD Working Group will choose, and we’ll look for comments, in this case from the At-Large community, and feedback on this.
The Working Group has some recommendations, but I would like to end with a set of questions. Which of those four models does the At-Large community feel is most appropriate? Are there any other models that we should consider but we missed? Which of the primary recommendations are feasible? Are there related recommendations that we should consider? And lastly, there’s also an issue of language tags: is there a need for the contact information to be in multiple languages and scripts?
So that’s all I have and with that I’d like to hear your questions.
Cheryl Langdon-Orr: Cheryl here. Thanks very much for that, Steve. I’m aware that not everyone is in the Adobe Connect room, so if you are in the Adobe Connect room you can use the usual method of waving your hand for discussion or interaction and Q&A with Steve. If you’re not in the Adobe Connect Room, Marc for example, let yourself be heard and we’ll start an order.
If everyone’s muted you’ve made it incredibly clear to us all, or nobody is wanting to interact. I know we have some noise on the line; can staff confirm we didn’t do a general mute for everyone did we, just to stop any of the noise on the line? Perhaps we need to have people do *7 to unmute themselves.
Gisella Gruber-White: We’re just making sure that all the lines have been unmuted.
Cheryl Langdon-Orr: Thank you, Gisella. Okay.
Gisella Gruber-White: The lines have been unmuted.
Cheryl Langdon-Orr: Thanks very much. Okay, Dave, you’re scrolling as is Dev typing, so if there’s some issues you’d like to discuss or questions you’d like to raise with Steve?
While we’re waiting for people to formulate their thoughts, Steve, I noticed in my very brief reading, and it was just a first reading through the four model recommendations, that for both model three and four the indication in the text was that there were considerable or reasonably large numbers in the Work Group that were raising concerns in those models, related to both cost to registrar and accuracy. Is there any sense coming from the Work Group about a least line of resistance or a preference for one of the four working models?
Steve Sheng: Not that I’m aware of. I think the approach is just put these four models on the ground and see what the community thinks which of the model should be used, and the Working Group will use that as a guide for its input on that. Perhaps it’s also useful for the community to provide us some guidelines. So for example, this is a classic policy question where we have to balance usability/accuracy on one hand but also cost, which is kind of the costs and benefits.
Perhaps it would be more helpful if I emphasize for the community to tell us what are the factors that are really important to you and that we should really put into consideration? Obviously accuracy is one important factor, but are there any other factors besides accuracy that the community thinks is really important?
Cheryl Langdon-Orr: Okay. Again, I don’t see people raising their hands but please do so. Steve, I wondered whether you wanted to raise some of the comparative points out of the report where you looked at the differences in costs or effects on the current registrars versus the IDN new registrars. Perhaps that might also help some of us formulate our thoughts.
Steve Sheng: Not specifically. So when we began we only had two models, and three and four was actually one model and that’s the key separation. And then immediately participants that are running registries and registrars raised the question of “If you’re going to do this manually this is going to be very costly.” So because of that discussion, the Working Group members suggested transliteration as there are already existing tools to do that.
We are also studying an automated translation system, so if the community knows of some of these we’d love to hear about one of these systems to bring down the cost while trying to achieve the same level or even more accuracy.
Cheryl Langdon-Orr: Has the Work Group looked at giving us some examples of accuracy? I mean I’m expecting Dev to come in from the Latin American and Caribbean point of view where they have used for some time some translation tools in their email system, and at least most of us are used to the consistent errors which come in on some of these automated tools.
I noticed you talk about recognized standards or performance standards put into some of these models, so it seems if there were going to be automated translation or transliteration – preferably translation tools from an accuracy point of view as I think you pointed out – would they have to be of a particular send or would the Work Group be looking at mandating particular types of tools? And if so would they be open source or proprietal? I think these are the types of questions that our community would probably want to grapple with as they’re putting their responses to the report together.
Dev, did you want to pick up on any of those points? Not wanting to put Steve in the hot seat to answer that rather than simply be able to reflect back to the Working Group some of our interests and concerns.
Steve Sheng: Cheryl, to answer your question- Go ahead, Dev.
Cheryl Langdon-Orr: Dev, go ahead.
Dev Anand Teelucksingh: This is Dev with a bad connection. Okay, can everybody hear me now?
Cheryl Langdon-Orr: Yep.
Dev Anand Teelucksingh: Okay. Apologies for the bad connection here. I did write in a few questions as you were going through this presentation so I just want to ask a different question. Regarding models three and four, isn’t it more the registry that is going to have the responsibility because it’s the registry that has to provide the actual IDN data and possibly the WHOIS data? And I apologize for making it sound very convoluted; I guess I’m still trying to organize my thoughts. That’s why I was trying to type it.
There’s the way that the registry operates the IDN, it does not impact on models three or models four. (interference) Can registrars create two different approaches and the IDN registry will still support it? I guess that’s what I’m trying to get at.
Cheryl Langdon-Orr: Steve, did you want to respond to that or do you want to take it as a question we will raise in our responses for public comment? Total silence, okay. Thanks for the question, Dev. I think that we make sure that we capture that question because it’s the type of thing that I think the IRD Working Group would very much like to grapple with.
Steve Sheng: Sorry, I was on mute.
Cheryl Langdon-Orr: Yes, Steve, go ahead.
Steve Sheng: Sorry. So the registrars’ versus registries’ responsibility, as far as the Working Group discussion goes the Working Group thinks this is the registrars’ responsibility because the registrars have the customer relationship with the actual applicants, and the registry would take whatever is the registrars’ information and put it into WHOIS. That’s my understanding, but perhaps this is very useful for the community to point out specifically. I know this question has been raised before as well.
Cheryl Langdon-Orr: Marc, I know you have an inability to get into the Adobe Connect room, but I wondered while you’re on the call, are there any particular concerns from your real work perspectives, particularly to the issue of is there a need for contact information to be in multiple languages and scripts – the last of the additional questions for the community consideration?
You might be muted, Marc. If you are you’ll need to press *7 to unmute.
Heidi Ullrich: Cheryl, this is Heidi. We’ve just heard that actually Marc is no longer on the call.
Cheryl Langdon-Orr: Okay, well that’s a pity. Perhaps we take that as a follow-up question to give him to look at particularly from a- I think we do need to look at the balance of the privacy and the use of the WHOIS information. It seems from my rather naïve perspective that the answer would be yes, there would be a need for contact information to me in multiple languages and scripts, but that need while wanting accuracy and predictability and (inaudible) of access, which is not always in keeping with other people’s views. Sebastien, go ahead.
Sebastien Bachollet: Thank you, Cheryl. Yeah, the question of accuracy of data is questioned here about IDN purposes, but it’s a more generic question on WHOIS and I guess it’s a subject of discussion in other working groups. How does this Working Group specifically want to solve these bigger problems or just the accuracy for a time other than ASCII script? Thank you.
Steve Sheng: That’s an excellent question. As far as the scope of the Working Group, it’s not tackling the accuracy, the general accuracy issue. Perhaps at least the Working Group does not want, with the introduction of internationalized registration data, that accuracy further suffers, at least. That one part is in the scope, but the general question of how to make WHOIS data more accurate, I think currently there are some… Let me think. There is an accuracy study being done and a report published. I think that issue is being carefully studied, but I think once we have that it would probably be that the Working Group has to adopt whatever guidelines come up. So the narrow response to your question is no.
Cheryl Langdon-Orr: A follow-up from you, Sebastien?
Sebastien Bachollet: No, in fact I don’t know if it’s a follow-up or another question. Do you work differently for thin and thick registries or does the Working Group just take one size will fit all? My question, it’s because one part of the model is with registrars, and yes, registrars have the data except for some registries who alter the data of the WHOIS. And is it something different for those that do two types of registries?
Steve Sheng: I don’t know. I don’t think this issue has been raised thus far, so I would look for the community to guide us in terms of the differences of the models and how that will apply to thick registries. So essentially we’re looking to see whether there is a difference in our models, either using a thin registry type or if we were to apply it to thick registries. So we’d like to seek guidance from the community on that.
Cheryl Langdon-Orr: Thanks for that, Steve. I noticed the comparative work on page 20 of the report where there’s a number of comments on models two, three, and four, which basically say the impact is unclear. This doesn’t help us a whole lot in weighing in on a user or an end user perspective on preference where we certainly would not want to be as a community recommending something where an impact albeit unclear at this stage had a serious or deleterious effect particularly on the rollout and uptake of the internationalized domain names in the new gTLD space or the ccTLD space.
And I guess one of the reasons I said those two things separately is because from an end user perspective of course, what we have heard a great deal of is that a continuity and predictability from a registrant or from an end user who wishes to access for legal purposes this type of WHOIS information in whatever script it is, assuming that it’s accurate and not bogus, should be able to do so and not see vast differences in the material that they have access to and the accuracy and methodology of getting that access whether or not they’re looking in a ccTLD environment or a new gTLD environment.
Are there studies planned? Is the Work Group intending to drill down further on these possible impacts or is it just where it said there is no impact or the impact is unclear we cross our fingers and hope for the best if we go down some of these pathways?
Steve Sheng: Those assessments are preliminary and informational only. Actually, I took a first pass, coming up with those categories to investigate. And so it’s very preliminary, that’s what I have to say. If I understand correctly you want us to do a drill down and see whether the impact is clear or not. That’s clear? Okay, we can do that.
What I understand is that the At-Large community’s needs and concerns are, the top concerns are accuracy and accessibility of the information? Is that the top two?
Cheryl Langdon-Orr: And also predictability of what is going to be found when we’re looking in an ASCII or a non-ASCII set of data.
Steve Sheng: Okay. So I would encourage those criteria to be permitted in the comments so that will guide the Working Group discussion.
Cheryl Langdon-Orr: I also note that there was some discussion in the paper on barriers to registrants, and obviously we would be concerned at lack of local language use being a barrier to registrants in new internationalized domain name spaces. It seems at least to my very biased point of view important that we do allow local language input from a registrant of preferably accurate information; and of course recognizing that under normal circumstances, whenever there is an increase in cost it tends to get passed onto the registrants always.
Steve Sheng: What you point out is a very good and important factor. I don’t know how to answer that.
Cheryl Langdon-Orr: No, no – we can’t expect a level playing field here, particularly when we look at differences between existing registries and registrar models and those who are yet to come in the new world, in the brave new world that we’re heading towards. (crosstalk)
I noticed you do have some additional slides. Are there any other materials you wanted to take us through?
Steve Sheng: I think that’s about it. Those are the main, the models are really what the Working Group is trying to get feedback on. And there’s another aspect of today’s WHOIS protocol, it’s not able to handle internationalized registration data. And there’s discussions on an enhancement or a replacement to the protocol. So that’s the feasibility part. But that’s just a policy question for now. I mean, I’m sorry, it is a very important policy question but not the one that the IRD Working Group is seeking feedback on.
Cheryl Langdon-Orr: Some of us again with our views would see forcing a policy review of some of the WHOIS policy as a very good thing, albeit somewhat traumatic for some. Steve, I’m wondering if the presentation you’ve put together, I’m noticing that at the moment the slide deck is not linked to our Confluence page. Are you able to make that slide deck available for the Community Call page so people can access it?
Steve Sheng: Sure, absolutely.
Cheryl Langdon-Orr: Okay, that would be very, very handy. I have to ask, are there any copies of this information, as in the slide deck, I know what the PDF options are in other languages but is there any language availability for the slide deck and presentation?
Steve Sheng: No, right now it’s only in English but I guess if there’s a community demand we can translate it into the six UN languages. I mean we translated the report into the six UN languages.
Cheryl Langdon-Orr: The report, yes.
Steve Sheng: The report is in the six UN languages, yes.
Cheryl Langdon-Orr: Yes, yes. And we appreciate that, believe me. Quite often those in the regional outreach work will find it rather handy if at least some of the key slides are in the languages that their communities have greater ease using.
Steve Sheng: Okay, we’ll take that.
Cheryl Langdon-Orr: Yeah, if Dev and the European and African regional leaders wish to approach you it sounds like it’s not going to fall on necessarily deaf ears. So some things are possible, coming from a person who doesn’t even speak or write English because she’s Australian.
Dev, you’ve raised a point in the chat (interference). Would you care to share that on the record? I’ve got an echo.
Dev Anand Teelucksingh: Okay, this is Dev.
Cheryl Langdon-Orr: Go ahead.
Dev Anand Teelucksingh: Hi, this is Dev, okay. The question- Can you hear me?
Cheryl Langdon-Orr: Yes.
Dev Anand Teelucksingh: Okay, good, sorry about that. Just a question looking at the report, regarding model one where registrants may have to enter the data twice, once in the (inaudible) script, maybe US-ASCII, and then optionally in a local script – how can differences be resolved? Because it seems to me that inaccuracies, you can have an inaccurate recording.
For example, a registrant enters their address as 123 Anywhere Street, and then in the local script it’s entered as 234 Anywhere Street. That’s my question. Is there any way that those differences could be resolved and did the Working Group think that therefore this model one may not be as suitable? That’s my question.
Steve Sheng: Model one is sort of today’s status quo. So I agree with you that this is a real case and I encourage that this particular question be submitted as an issue for the IRD Working Group to consider. I mean I think today it’s possible so it’s probably no less accurate than before.
Cheryl Langdon-Orr: I was rather hoping we might be able to pitch it at a higher aim than a continuation of the lack of accuracy we already suffer in the ASCII world.
Steve Sheng: So to address that, registrants… So if today a registrant for privacy reasons submits bogus information, going forward in the internationalized era that is still going to happen, right? So even if not in this case, if a registrant just enters WHOIS information that’s completely bogus for the right reasons, for his or her privacy protection, I don’t see that helps or how internationalization will make that worse. If registrants want to do that today they can do that. Of course I’m just thinking through this issue.
So I guess the question that I have personally then is if the community can tell us where the situation could be worsened because of the introduction of internationalized registration data. Those scenarios would be very important to us.
Cheryl Langdon-Orr: Thanks for that, Steve. Can I make a call for anyone who wants to make any further questions or comments at this point?
Steve Sheng: I’m sorry – can you what?
Cheryl Langdon-Orr: I’m just calling for anyone else who wants to raise any further questions or problems at this point. I’m not seeing anybody in the AC room or hearing anyone saying “Hey, what about me?” Oh, Sebastien is typing. I think this is going to be one of those-
Operator: Hello?
Cheryl Langdon-Orr: Yes, go ahead.
Operator: This is the Adigo operator. Can I speak to the chairperson of this conference?
Olivier Crépin-Leblond: Staff, can you help me? Matthias, are you there?
Matthias Langenegger: Yes, but I didn’t understand what this person was saying. I’m trying to find out where this person is calling from.
Cheryl Langdon-Orr: Thank you. I thought they said they were the Adigo operator, which is genuinely scary if that’s the case. Matthias, while I have you on live voice unmuted, this is going to be a call where I think some of our discussion, unfortunately some of the audio quality hasn’t been particularly good but I think we need to battle through that.
I think because of the audio quality not being particularly good, this might be one worthwhile to see if we can get some transcript material out of and also capture the questions raised in the chat. There’s been I think some very worthwhile discussion and issues raised which can form the foundation for the response to the public comment and questions for community consideration. But to make the job easier as the regional leadership reach out to their At-Large structures, I’d rather them work with text I think from this call if at all possible, rather than trying to play through some 50-odd minutes of audio file.
So I’ll leave that with you to make sure someone picks that up as an action item.
Matthias Langenegger: Sure, yeah. We’ll ask for transcription.
Cheryl Langdon-Orr: Thank you, I think it might just help. Steve, I have no idea what that last intervention was, whether it was from the Adigo operator or not, but if we can find out what it is and if it’s something that’s relevant we will include it. I take it that we’ll have a copy of your slide deck on the Confluence meeting page, and you will be more than happy if any of our regional leaders email you directly or through Matthias and Heidi and Gisella. Is that okay if you have some follow-up work to do?
Steve Sheng: Yep. Thank you. And I will also enter my contact information. The best way to submit comments is through the public comments but I will enter my contact information as well, and the IRD Working Group’s email address.
Cheryl Langdon-Orr: Oh excellent, very good. Greatly appreciated. Well Steve as ever, those of us who are on the call have thoroughly enjoyed your presentation. Sorry about some of the audio quality challenges. I don’t know whether you were hearing some of the background noise during your presentation but I know the staff were working fairly frantically to make sure they could make it better, and it was about as good as it could get in some situations. But if we capture what we can in text here and we will have more than just the questions and comments raised here, and in our responses that we’ll be getting in before… Is it the 14th of March? Is that the date just to double check?
Steve Sheng: Yes, yes, the 14th of March.
Cheryl Langdon-Orr: Terrific, okay. Being a mid-month event it may be that what comes into- And I gather you’re the staff person for this public comment material as well? You’ll be doing the follow-up pieces after the PC’s close? Is that the case, Steve?
Steve Sheng: You mean I’ll do a summary of the public comments?
Cheryl Langdon-Orr: Yeah, is it you doing the summary of public comments?
Steve Sheng: Yeah, it’ll probably be me but we have… Yeah, it’ll probably be me or Julie.
Cheryl Langdon-Orr: Okay. The reason I was asking, Steve, is because we will be putting in our public comments I’m sure it’ll be a fully ratified – i.e., ALAC voted – set of comments, but whether it’ll be coming in as drafts on that day or if it will be fully ratified it won’t make any difference if they come in under vote. But you will be notified so this shouldn’t hold up your next steps at all. But you will be fully notified when they become a formally voted on set of comments and therefore become a statement.
So providing you can live with that quirky way of us doing things we’ll do our best to get some of the community considerations well and thoroughly discussed and sent back to your Working Group to chew over. And they certainly do have a lot more to chew over by the looks of things.
Thanks, Steve. We all appreciate the work you and the Work Group have done. Thank you to all of the participants on today’s call, and to catch the phrase where people are saying good morning, good afternoon and good day, thank you all for your contributions and for joining this call. Bye for now.
Steve Sheng: Thank you.
[End of Transcript]
