*** Disclosure: The following is the output resulting from
transcribing an audio file into a word/text document. Although the
transcription is largely accurate, in some cases may be incomplete or
inaccurate due to inaudible passages and grammatical corrections. It
is posted as an aid to the original audio file, but should not be
treated as an authoritative record.***


IDN Variant TLD
Monday, 20 June 2011
ICANN Meeting - Singapore


>> Ladies and gentlemen, we're about to begin our program.  It is the
IDN variant TLD.

If you move forward in your seats, it would be much easier for our
panelists and much more interesting for you.

Thank you.

>>KURT PRITZ:   Hi.  Welcome, everybody, and thank you for coming.

Could you guys in the middle sit down or....

Hi.  I'm Kurt Pritz.  I am the senior vice president of services at
ICANN.

Thanks very much for coming to what is this session about very
important work that ICANN is undertaking with a renewed vigor with the
idea of delegating variant TLDs.

It's very important now that IDNs have been delegated that we continue
to work in order to provide a good user experience for those who use
IDNs so that they can fulfill their potential in broadening the
participation of the use of the Internet across the world.

Variants, as the people in this room know, are a very difficult issue.
Variant TLDs contain characters.  Because those characters represent
the same character in an alphabetic set, they can be very confusing.
And so we have options right now.  Either delegating variant TLDs that
might result in user confusion and a poor user experience, or not
delegating variant TLDs, and essentially disenfranchising a sector of
that community that uses that alphabetic set.

So it's really important that we do this work to resolve that
conundrum so that we can delegate new TLDs that are variants and
guarantee a good user experience.

Now, there's been a lot of work on this in the past, several working
groups.  And the experiences from those working groups have taught you
as few things.

One is that when we investigate issues associated with variant TLDs,
we make good headway, say in one script set, say Chinese, but then we
think that really won't work for Arabic, and the discussion of trying
to develop one set of issues and one set of solutions across all
scripts has been difficult and problematic.  And so members of the
ICANN community have hit on a solution that, or a potential for
solutions that involve the conducting of six separate case studies,
one each for different scripts, and identifying the issues solely
associated with those scripts.  And that way, work can proceed on each
script, a set of issues developed, and then later on, solutions
developed for those scripts.

We'll find that some issues, of course, are across all scripts but
some issues are script specific.  So we think this, conducting six
different case studies for six different scripts, is a very good way
to go about investigating what the issues are.

ICANN's been very fortunate in attempting to recruit participants for
these studies.  I think we presently have 65 participants from 30
countries staffing these six different panels.  And we also have the
contributions of several host occurs that are contributing meeting
space or a travel budget or Secretariat support, and other means for
supporting this project.

So in a very short period of time, we have a broad reach of
participation across the community.  And this meeting really
represents the start of substantive work on that.  Each one of the
case-study teams has already met in developing a set of definitions
and carrying on the future work, and each has made separate plans for
how they are going to carry that work forward.

So it's really my pleasure to introduce the people behind me.  I'm
flattered is exactly the right word.  I'm so flattered that this set
of people have come to work with ICANN.  And when I say "ICANN," I
mean the big ICANN community, not the bunch of staff members, to help
solve this problem.

And so I'm going to not introduce them all.  I am going to, instead,
introduce Dr. Dennis Jennings who is a former board member and is
performing a lot of work in coordinating this group and providing
leadership in getting the right people on board, organizing the work,
and attracting host organizations such as we've had.

So I want to thank you very much for coming, for your interest in this
topic, and I want to start by thanking Dennis for taking over for me
and starting the presentation.

>>DENNIS JENNINGS:   Thank you, Kurt.  Thank you very much, indeed.

If we could have -- Well, we have this first slide up.  Could we go on
to the next slide, please.

So why this project?  Well, it's been a longstanding request from many
people in the community that we address this issue.  In fact, it's
been a longstanding request that we solve this issue.

But when we look at it, we discover that there are many unanswered
questions.  And so the board instructed the CEO and the staff to
develop an issues report on the subject.

What do we mean by an "issues report"?  Basically we mean a statement
of requirements or a statement of the problem that we're trying to
address.

Now, a lot of people think that the problem is easily stated, and it
usually is.  It's "the problem is easy.  Solve it for me."

And it turns out when you look at the whole variant issues, even the
definition of variants is not agreed across the set of scripts.

So this project is focused on issues, not on solutions.

So where are we with IDN variant TLDs today?  Well, in the new program
just announced -- and congratulations to the board and to Kurt on that
achievement -- applicants may declare variant strings for the TLD in
its application.  Some work needs to be done, I think, in that area in
the Applicant Guidebook, but that's the idea.  But no variant TLD
strings are going to be delegated until variant management solutions
are developed and implemented.  And the first step in this is to
understand what the issues are, what the statement of requirements is,
what the problem is so that we can then move on to develop solutions
for IDN variants.

So here's the scope of the project.  First thing is to develop or
create a glossary of terms that's been vetted with the communities.
And that turns out to be an interesting challenge, which I'll talk
about in a moment, because the terms are not precisely defined.  They
mean different things to different people, and it turns out they mean
different things in different scripts.

So that's the first thing, to create a glossary of terms that is
consistent across the scripts that we're trying to address, and then
to identify the challenges or the requirements, the problem statement
based on linguistic accuracy, technical feasibility, usability,
accessibility, security and stability, and so on.

So the first proposal or the final proposal for this was published on
the 20th of April.  The initial proposal was published for this study.
There was -- published for public comment.  We got comment back.
We're very grateful to the community for, given all the other things
that were going on, to actually look at this and to give us comment.

And as a result of that, of the comments, we added a Greek case study.
We had originally proposed only five studies:  Arabic, Cyrillic,
Latin, Indic and Chinese.  As a result of the public comments, we
added a Greek case study, and we also refined what we had called the
Indic case study, which is a misnomer because there isn't an Indic
script, to be the Devanagari case study, a much tighter -- a single
script, a much tighter definition.

So we got a lot of positive feedback from the community, and that's
the final proposal that was published.

So the next step was we published a call for volunteers for the six
case studies.  We received over 70 applications across the six case
study teams, and we did a lot of work to clarify those issues, and
we've completed and published the selection of the team members.  And
um see that information later on in this presentation.

And we selected the case study coordinators, the key people who will
take responsibility for these six case studies, major projects in
themselves within the project that we're undertaking.

The teams are comprised of experts in a variety of relevant areas, as
you can read on the slide up there.  The DNS, operations, linguistics,
security, policy, and, indeed, the IDNA protocol.  They are led by the
case study team coordinator, and they are going to have regular
conferences, face-to-face meetings to do the work.  And we in ICANN
will provide support to that.

In addition, we have bolstered our own team with some external
experts, and there are two named there.  Andrew Sullivan, who is a DNS
and IDNA expert, who provides a lot of help on the technical side.
And Andrew isn't with us.  He is in Canada at the moment.  But he did
participate by phone on the various meetings we've had so far.  And
Nicholas Ostler, who is a linguistic expert who is with us and who has
identified himself by raising his hand so that you know who he is.

In addition, we have selected, and we're very pleased to have had a
number of proposals for host organizations.  As Kurt said, they will
provide a variety of support from meeting space through to
administrative support and, in some cases, even financial and travel
support for the case study team, varying from case to case.  And we're
very pleased and we announced this, we published this the other day,
that ICT Qatar has been selected, volunteer has been selected as the
Arabic host organization.  CDNC, the Chinese Domain Name Consortium,
for the Chinese case study.  UNESCO in Paris for the Cyrillic case
study.  The Department of Information Technology, the DIT, in India,
and their department CDAC and NIC SI for the Devanagari case study.
ICS-FORTH or FORTH-ICS in Crete for the Greek case study.  And dot SE
in Sweden or the Internet Infrastructure Foundation for the Latin case
study.  And we are really pleased to have these organizations roll in
behind these studies.

As you saw from the scope of work, the first thing to do is create a
glossary of terms that everybody could agree with.  And we have
developed -- the team has developed with a lot of help from Andrew and
from Nicholas a definitions document defining the terms being used.
And that has been circulated and will be published.

It's an initial document.  We felt very pleased with it.  We thought
that we had this kind of nailed down, and it's interesting that much
of the discussions on Saturday among the case study teams have
indicated to us that there's quite a bit more work to be done on
definitions, and the definitions may vary from case to case.

But what we're trying to do is to make sure that we have a single set
of definitions, and there may be slightly different definitions for
different cases, but we are going to use words with a clear meaning.
We want to have a clear glossary of terms.  And if there is a slightly
different definition for one case, we're going to use a different term
for it so that it's clear, so that when we use terms, and hopefully
when the community use terms, we'll all understand precisely what that
term means.  And that is certainly not the case at the moment, and I
can tell you there is a lot of discussion you may have heard over the
last couple of days where the word variant is bundled -- bundled -- is
bandied around in a variety of fashions that are completely
inconsistent with each other.

The case study team feedback is required.  That has been started.  We
started with two types of definitions from existing references and new
definitions, and this is very much a work in progress, as I said.

So we asked the case study teams to use the terms as defined.  I
already commented on that.  Really, I've covered this slide in a
variety of ways in what I've just said.

And the goal of this and the restrictions that we're trying to apply
is to minimize the difficulty later on during the harmonization step.
I repeat, so that when we use a word, that everybody knows precisely
what we mean by that word.

So how do you guys in the community stay informed?  Well, there's a
mailing list that you can sign up for where there will be information
posted regularly, and where you can participate.  And there will be a
community Wiki that you can sign up for, and the URL is there.

And the work has started.  The case study teams met -- what day is
today?  Today is Monday.  On Saturday.  I am already lost and it's
only the first formal day of the ICANN meeting.

On Saturday, we had a whole day meeting with the six case study teams,
a couple of plenary sessions to work together, and individual
sessions.  And I'm delighted to say that all the teams got to work and
did a lot of work, a lot of hard work there.  I'm very impressed with
the amount of effort that the teams are putting in.

And the case study coordinator will present here a brief snippet of
the work and the ongoing work and the plans for each of the case study
teams.

The work has only started.  There will be lots of telephone
conferences, there will be meetings.  The schedules aren't yet -- or
schedules as the Americans say -- aren't yet finalized but they will
be published as the case study teams develop them.

And the intent is, and we're holding to this, if we can, and it's up
to the case study teams to try to deliver to this, that each case
study team will develop an issues report by the 30th of September this
year.  Not the 30th of September next year.  The 30th of September
2011, this year.

That's challenging, particularly in the northern hemisphere where, as
you know, particularly in Europe, lots of people go on holidays.  It's
also challenging in the Arabic script case study because Ramadan is in
August this year, and I'm sure there are other challenges, but I know
every case study is going to try to deliver the report by the 30th of
September.

And then our goal for the final issues report where case study
coordinators will get together, probably in Senegal, in Dakar in
Senegald, to try and see whether we can have a harmonized report or
whether we have six individual reports with a report that highlights
those things where there is agreement across the reports.  That's due
by the 15th of December.

So now let's hear from the individual case studies on some of the work
that's done.

And is Sarmad Hussain here with us?  Yes, indeed.  Sarmad, would you
like to take the microphone and given us a brief update.

>>SARMAD HUSSAIN:  Hello, everybody, and thank you, Dennis.

I'd start -- So what I'm going to do is very briefly overview,
introduce my team, the team for Arabic script case.  It comprises of
some very experienced people who have been working on Arabic script
and Arabic script-based languages for some time.  We have Abdulaziz
(saying name )from Saudi Arabia,

(Listing names).

All these people bring a lot of experience in this area on the table,
and that makes the work easy in a way because they are very familiar
with the issues which exist with Arabic script.  But also, you know,
we understand the challenge because we have to actually close this
work in a very short time.

I'll just introduce some of the challenges which are relevant.

So what I've done here is put the IDN ccTLD string for Pakistan at the
top.  And you can see that both strings which are shown are exactly
the same visually, even though they use different code points from
Unicode internally.

And just to highlight the significance of why variant TLDs may be
required for Arabic script.  The second example is just a different
word, it's not a TLD, but it highlights a different kind of problem,
for example, where it's not that the different characters look the
same way, but a combination of a combining mark -- excuse me -- with a
letter looks exactly the same way as a letter which already is
separately encoded in Unicode.

And some of these cases, for example, are dealt with normalization in
Unicode, but there are many such cases in Arabic script which are
actually not dealt through Unicode normalization.

So just to give you very briefly some examples of the issues we are
dealing with as far as TLDs are concerned.

Next slide, please.

And here is a very brief but aggressive schedule which we are going to
follow.  And I must thank Baher for putting this slide together.  And
just to take you through this, we are actually going to take two
different activities, the two different activities by the team.  And
I'll start with the issues list.  So we are going to do two separate
documents.  We are going to work on the issues and work on
definitions, as Dennis pointed out as well.

We have already identified a first list of issues which -- yesterday,
and we plan to go through and expand our case as much as possible
through the e-mail list we have and try to close -- agree and close on
the issues by the end of the week this week.

And then based on those issues, we are going to draft a document with
clearly -- which clearly explains those issues with examples.  And
what we are trying to do is not just come up with some representative
examples but actually come up with an exhaustive list of those
examples for each of those issues.

And then eventually finalize -- once the document is prepared,
finalize the document based on internal discussions and then hand that
over to the IDN VIP team.  In parallel, what we're going to do, though
we initially started looking at the definitions document which was
given to us by ICANN, very quickly on the first day we realized that
it actually may not be possible for us to close the definitions
without really looking at issues in more detail first.

So what we're now going to do is finalize the issues first.  And then,
once the issues are taken care of, we'll come back to the definitions
document.  And first try to define the definitions from our context.
And then look at how ICANN is defining these items or glossary items
and see where we match and where we don't match, see what the
differences are and whether they can be incorporated into the larger
document.

So that's what the brief plan is.  We'll first do the document, issues
document, then do the definitions document and then, hopefully, hand
both things over to ICANN by the end of September.  So thank you.

 >>DENNIS JENNINGS:  Thank you very much, indeed.  A very impressive
 piece of work.  So you're well up and running.  And thank you for
 that.

I'd like to move on to the Chinese case study.  Is professor Xiaodong
Lee here?  Xiaodong, yes.  Would you give us the highlights of the
work so far?

 >>XIAODONG LEE:  This is Xiaodong Lee.  I think I am the coordinator
 of the Chinese case study team.  Can we move to the next slide?
 Yeah.

So I want to give a very brief introduction about the case study team.
You know, I'm very happy that we have the local host Chinese domain
name.  The CDNC is -- I've been working for Chinese domain name for
many 10 years.  I think there are so many memories from the Chinese-
speaking area and also some from the Europe and Americans.  They have
so many discussions.  So I think if we have the CDNC be the local
host, it would be smooth, some case study work.  Next slide.

I'm going to just give you a very brief member highlight.   You know,
as the CDNC is local host, in the team members we also have the people
from CDNC, the Professor Seng, Professor James Seng is the chairman of
NIC and is also cochair of CDNC.  And we also have Dr. Zhang.  Also on
behalf of the doc CNIC.  And, you know, I think you have to remember
that the speech (saying names) is the IDN working group chair.  And
James Seng is also the main contributor to the registry.  Myself is
the also contributor of the 4713.  Also the fundamental document for
the Chinese issues.  And, you know, I also cochaired the -- I'm
informal cochair of the IETF internationalized email address working
group.  But I'm very happy that we also have Professor Zhang is the
main contributor of the Unicode CDCJK.  And also the main contributor
of the Chinese variant table.  It was published in IANA.  Lists
variant table have been adopted by so many registries which provided
the Chinese domain name registration.  I think next slide?

Also I'm very happy to have the team member Joseph Ye and also (saying
name)  Those two people are also the current IETF working group
cochairs.  Joseph Ye is co-chair of the international e-mail address.
And Yoneya is also the cochair of the IETF working group.  So these
two working groups is the current active working groups which are
involved in the internationalized issues, IETF.

So we also have the representative from registries from dot CN, dot
HK, dot TW, dot com, dot Asia.  I think some of them operate the IDN
TLDs.  And we also have the community person from the communities from
APTRD and also from Chinese domain name user in the Asian Pacific
region.  This case study team is distributed.  We have team members
from Asia Pacific region, but we also have members from Europe and
American.  So I think that this case study team will work together to
finalize the issue report.

I want to give some information about the past two days discussion
meeting.  At the request by ICANN, we have finalized the issue part of
Chinese case study in the end of September.  But we have a consensus
in that case study team which should review, finalize the question
list.  And the -- in the end of July.  And try to initial the issue
report in August.

And the first question of the issue probably should be released in the
early of September.  So it will give us much time to discuss and to
get the review from other expert.  I'm going to -- except for the team
member of this case study, we also want to conduct the other expert
from IETF or from other organizations to make sure the issue report is
accurate and right.  I think, you know, the Chinese variant issue is
not a fresh new issue.  The Chinese community has been working on the
Chinese variant issues for over 10 years.  Even in the very beginning
of the IDN internationalized working group in IETF, the variant issue
is a very important issue.

So many discussions in IDN working group in IETF, how to do the symbol
for Chinese and also other kind of variant issues for Chinese
language.

In the past 10 years since it was published, it's not a very perfect
solution.  But it is workable and acceptable by he Chinese community.
And I also want to mention that now IETF is also working on some new
solutions for the variant issues.  It's not only for Chinese case.
It's also maybe acceptable by other language.  I also see that it's
involved in this room.  So -- but I want to clarify that that maybe
two states now for the Chinese variant issue, it maybe found an
acceptable solution and ways that the issue report.

But in the next state maybe we can find more perfect solution in the
IETF or other organizations.  But the most important work for this
case study team is to clarify the issues and finalize the issue
product.  This issue will be useful for the ICANN community.  I think
that all the members of this case study team work very hard and try
our best to finish this work.  Thank you.

>>DENNIS JENNINGS:  Xiaodong, thank you very much indeed.  Excellent,
and excellent to hear you have a plan to beat the deadline and get
comment out early.

Can I ask Alexey Sozonov who is coordinator for the Cyrillic to give
an update?

 >>ALEXEY SOZONOV:  Okay.  Brief update of our Cyrillic case study
 group is we had a meeting and we understood that here is the scope of
 the broad definitions which are for all languages.  And which need to
 be boiled down to the definition precise for Cyrillic languages and
 Cyrillic script.  And just from the first meeting we acknowledged
 that we could not cover all Cyrillic scripts since it's too many
 languages and small languages.  So we'll cover languages which is in
 the scope of the members of our group.  And we'll have two experts,
 Sergei and (saying names)  They are current authors of Cyrillic
 script 5992.  And here is the list of the members of our group.
 Next, please.  We also understood that we need broader community
 input during the discussion period on the Wiki to bring the issues
 that might be overlooked by us.  As well we think it's a good idea
 will be to have communications with other coordinators and to share
 the working program because sometimes people look to things from
 different perspective and they bring up issues which was overlooked.

We decided to meet once a week.  And we might have a necessity to have
a face-to-face meeting in the end of that work.  And we are happy that
our host organization is UNESCO.  We didn't know about that.

And next slide.  Here it is.

And, since we've been learning the current definitions, we understood
that they were created for different languages and different cases.
And most of them are much more complicated than Cyrillic requires.  So
we need to work on it together with an expert.

And thank you.  This is all update.

 >>DENNIS JENNINGS:  Thank you very much indeed.  Excellent.  Could I
 ask Dr. Govind to speak about Devanagari -- I think I will get this
 word correctly pronounced one of these days -- to talk about the case
 studies.  Thank you.

>>DR. GOVIND:  Thank you.  Thank you for the opportunity to elaborate
about the Devanagari, which supports major languages of the country.
And thank you for the opportunity to get us into a team, which will be
going into the various aspects of this IDN TLD variation.  So our team
will be consisting of ... (listing names).

I will be taking time to elaborate what these persons like the
community representatives will be that the Devanagari script will be
(saying name.)  The linguist will be consisting of Dr. Rajiv and Mrs.
Shashi (saying names.)  They have more than two decades of experience
in the linguistic aspects of the Devanagari.  Then we have
registry/registrar operations.  We have (saying names) and Kumar who
belong to NIC SI and Afilias.  They are technical service provider.
They will be assisting us in the registry/registrar operations in the
IDN TLD variations for Devanagari script.  Then we have a policy part
which will be done by (saying names.)

We have taken from the diverse backgrounds, diverse agencies who will
be assisting for this activity of IDN TLD.

Then we have (saying names) and James Galvin from Afilias who will be
assisting us.  Then we have security and stability.  Again, James
Galvin and Amardeep, who is one of the third operations in the
country.  Then we have one of the directors in that.  So, overall,
this team comes from a rich background and a diverse knowledge of the
aspects of the computers and the security aspects, variant aspect, and
the script variation aspect.  So all this, and community expertise is
also there.  So we expect that this composite team will help us to
come out with a good study on this Devanagari script.  Next, please.

Some of the complexities of the script like Devanagari I will
elaborate.  I don't have the slide here.  But it is like L and 1 in
the Latin example a homographic variant.  Like yesterday we talked
about dra and dra.  These two look like similar looking scripts.  So
the definitions need to be suitably worded to accommodate the nuances
of complex script like Devanagari.  So that is what we need to see is
how the various complexities of the variant homographic variations in
the Devanagari, how can we accommodate that in the definitions, which
looks like more generic definition and which can take account of that
variation.  The word "register" will be changed to "located" and
"activated" be changed to "delegated."  That's what we discussed the
last two days.  And the character variant definition needs to be
redefined preferrable variant dominant uses be considered.  And 1:1
mapping of code in this variant is not the case in Devanagari.
However, multiple code point sequence may have variance.  And the
language variant table is given what needs to be accommodated this
kind of requirement.  This will be elaborated in our discussions with
the team from ICANN and give them the examples how this kind of -- why
this makes sense Devanagari is not 1:1.  It is one to many and many to
one kind of mapping.  We need to consider this aspect very carefully
while defining these definitions and putting that in a proper
structure.

The case study group will produce a table containing an explanation of
what they understand on the use definition and provide examples.  That
is what we discussed at each of the Arabic, Cyrillic, and the Chinese
groups.  If they can come up with a kind of table, what they
understand from these definitions, which is given by ICANN on each of
the IDN variant, and how each definition is supported by the examples.
So that will help us to go more into the commonality between the
various definitions and the examples of each script which we are
studying here.  The same should be shared with all the study group
members.  That is also we discussed.

The coming to the timelines, what we discussed, since it's a complex
script.  So we -- we will have a weekly meeting teleconferencing
planning to start from 27th of June, immediately after reaching our
entry.  And so that, on weekly conferences, we discuss the nuances of
various complex structure of the Devanagari script and how we define
the definitions and the examples.  Then to come out with a overall
structure of the Indian Devanagari script languages, we are organizing
a conference on 6 of July in New Delhi with specific references to new
gTLDs where we are inviting all the registrars, Internet community,
the government officials, the industries, so that they also know.  And
we're inviting representative from ICANN also so that they can see how
the Devanagari script and the Indian languages have a complexity how
we can accommodate it.  Face-to-face meeting planned in Pune.  In one
of the meetings since conference calls, we discussed how we can have a
face-to-face meeting in Pune.  And dates have also been finalized, the
21st and 22nd of July, where we can have a person from ICANN, IDN
group, and the various stakeholders of the IDN in India.  And we're
organizing this meeting in Pune where the premiere institution of the
language technology, CDAC, is based.  And then we intend to finish
this job by 30th of September, which is the timeline for completing
that duty.  Thank you.

 >>DENNIS JENNINGS:  Thank you very much indeed.  An extraordinarily
 complex case.  And very impressed that so much work has already got
 started and so much planning done.  I'd like now to ask Panagiotis
 Papaspiliopoulos to speak on the case study on behalf of Vaggelis
 Segredakis.  Vaggelis, unfortunately, was injured in a car crash just
 after he accepted.  I hope there's no association between accepting
 to be a case study coordinator and a car crash.  But, unfortunately,
 he appears to be quite badly injured.  And we send him all our very
 best wishes.  So, Panagiotis, can you give us an update on where we
 with are the Greek case study.

 >>PANAGIOTIS PAPASPILOULOS:  Thank you, Dennis.  Yes.  I'll speak on
 behalf of Vaggelis who had a serious accident two or three days
 before he was going to take a plane to come here to Singapore.  Can I
 present you, please, the members of the Greek team?

So Vaggelis is our coordinator.  And he's based in Crete, where it's
the registry of dot CR domain names.  And they have a lot of
experience there in handling the Greek letters.  We have -- for your
information, we have introduced the Greek letters and the dot CR for
the second level domain six years now.  And the variant case was
examined technically and unilaterally at that time.  And so we believe
that we can offer and transfer our experience and our problematic to
the ICANN project.

And next member is Asimina (saying name) member of the regulator.  And
Catherine Tsapikidou is working for the ministry of foreign affairs.
Evangelos (saying name) is an expert working for the Greek
civilization organization and has involved many times in formulating
relevant ISO standards, international standards, of course.  And one
of them is the Unicode standard.

Fotia Panayiotou is a member of the dot CY registry.  But CY is for
Cyprus.  Obviously, since there are two countries with the official
language the Greek, Greece and Cyprus, here help would be valuable in
order to have a common approach and common solution to the problems
that might arise.  George Papavlou is a member of the commission, of
the telecommunication.  It's the regulator.  And he used to work for
the commission, the European Commission.  And maybe some of you may
know him well, especially some people who follow ICANN meetings a long
time.  His experience is expected to be very valuable in this project.

Giannis Papaioannou is a member of the heliographical service of the
Navy.  It's a  linguistic expert and also an expert in geographical
names.  And he's also participating in several meetings around the
world, and he's representing Greece in the relevant fora around the
world.  So we expect his help and his experience to be valuable as
well.

And it's myself.  And I am a telecommunication expert.  I'm working
for the ministry of infrastructure transport and networks for the
general secretariat of communications.  So I'm nine years now.  And we
had cooperated with Vaggelis during the launch of the Greek -- for the
second level domain and the dot CR six years ago.  And I'm following
ICANN almost nine years now.  And I'm the GAC representative of
Greece.  So this is our team.  And we hope to do our best to present
something that this -- we all want to be useful, not only to the ICANN
staff but also to the Internet community.  And we have realized that,
even if Greek is a language that not too much people speak, it's a
language that too many people learn and maybe they want to use on the
Internet as well.  So, in coming to the project now, we would like
first of all, to define and identify character set since Greek
characters can be found in monotonical and polytonical systems.  And
it's set to tables in the Unicode standard that covers the set of
Greek characters.  So we have identified the set.  For your
information, all the characters are supported for the second level
domain and the dot CR.  And, after that, we want to present the
variant issue and to provide to the definitions and to -- for the
verification of variants to provide as many examples as it is
necessary in order to identify the problems.  Just a small brief, the
variant can be two kinds of variants.  One kind of the variants could
be if you put the tone on a different vowel.  So the word in capitals
looks the same but, actually, it's not the same.  And the capital is
the problem.  Because, otherwise, the encoding is different.  And then
the other case is when you have the same two words that look alike.
In this case, we have realized from the very beginning that the --
this could be a field that we should cooperate also with the Latin and
the Cyrillic team, since many of the Greek letters are also in the
Latin alphabet and the Cyrillic alphabet.  We believe that we have
experience.  And we are optimistic, even if our coordinator has a
serious health problem, that we will meet the deadline and we will
provide ICANN the full identification of the variant issue.  And,
since we have, as I told you before, some regulatory experience, even
if it's not in the scope of the project of the states, maybe you could
provide some information, some proposals for this as well.  Thank you
very much.

>>DENNIS JENNINGS:   Well, thank you very much, indeed.

Can I call on Jothan Frakes, now, who is the case study coordinator
for the Latin case.

>>JOTHAN FRAKES:   Thank you very much, Dennis.

My name is Jothan Frakes, and I am a volunteer with the Mozilla
project working with the application developer community on
assimilating IDN into use for the general user community.

And it's a privilege to have the opportunity to serve working with a
very good volunteer team.

If we could have the next slide, please.

Working with Dr. Andrzej Bartosiewicz,

(Listing names).

We also have staff support from Kim Davies, helping us work together
on some of the issues that exist in documenting those, capturing
those, as well as definitions and around those with respect to the
Latin script.

We have determined, like many of the others who have spoken here today
have talked about the definitions in and around the use of the term
variant, what those mean within each of the language study teams.  And
we're working hard on defining what those are.  And in fact, trying to
determine some of those terms -- terminology and variants that do or
don't apply, but at least documenting those.

So the people who finally are the consumers of the end result
documentation that we do are able to make well-informed decisions
about the next steps.

We're working with a language set -- excuse me, a script set that has
a fairly good exposure in a very simplified form in that A through Z
have been available for well over 25 years now in the DNS system.

We're working to capture different elements of user experience and
expectation and how those feed into the use of Latin characters as we
extend end into diacritic marks and ligatures, et cetera, as we expand
the Latin character sets that would be available in the Domain Name
System at the top-level domain.

We're trying to factor in things like what the user experience is with
respect to their QWERTY keyboards, Q-W-E-R-T-Y for the transcriptor,
as well as AZERTY, A-Z-E-R-T-Y.  And these different layouts mean
different things in different countries, in different uses.

We are also taking a look at the display of characters and how that
expectation works within applications and fonts for what the user
experience is.

An example of the user experience would be that in the current DNS
system that whether you type a capital letter A or a lowercase letter
A, the DNS system treats those identically.

And we're working with some character sets and some language character
repertoires that are fairly well structured using the Latin code sets
-- or, excuse me, scripts from within the Unicode definitions.

It's our objective to help document what we consider to be variants.
And there are a variety of variant classifications, such as
homographic characters as well as homophone strings, there's
homographic strings as well as orthographic variants, and a number of
other variants.  So we're working to define variant so that we have a
common vocabulary amongst our team that we can articulate to the other
working groups.

It we have also determined that we will have some exposure and need to
interact with other teams such as the gentlemen to my right and left,
Panos and Alexey, with the Greek and Cyrillic teams as well, because
there is some overlap in those character sets as well.

As far as the schedule goes, our team will be working with a biweekly
schedule call.  We are making a determination if the necessity of a
face-to-face meeting will be there.  We hope to ultimately make that
determination by the end of July.

Ultimately, we are pleased to have a very good team composition.
We're already well under way with assignments, and we hope to work
within the September 30th deadline with no issues.

And that is my update.

Thank you.

>>DENNIS JENNINGS:   Jothan, thank you very much, indeed.

I realize that I didn't introduce a couple of key people from the
project team, and I would like to do that now, and these are the case
study liaisons.  They are the staff on the team who are the key
contact points and liaisons with the case study coordinators, and they
are the people that you should know about on our side.

And for the Arabic case study, Baher Esmat is the liaison.  Baher is
in the audience there.  For the Chinese, Steve Sheng is the
coordinator, who is on the panel up here.  For the Cyrillic, Francisco
Arias, who is beside me here, is the coordinator.  From the Devanagari
script, Naela Sarras is the liaison.  For the Greek, Julie Hedlund,
who is on the panel up here.  And for the Latin, Kim Davies is -- is
Kim here with us?

>>KURT PRITZ:   Kim is unfortunately across town at a DNSSEC event.

>>DENNIS JENNINGS:   That's right.  He is at a DNSSEC signing ceremony
but I think Kim is well-known to you all.

I am going to open the microphones to questions from the floor in a
moment, but before I do that, just a couple of things.

When I started working on this project, I, like many other people,
felt that this was a technical problem, and it should be thrown over
the wall and a technical solution should be found.

I've been rudely abused by my technical friends and colleagues that,
in fact, it's not a technical problem.  And there is no technical
solution.  And if there were a technical solution, it would take a
long time to implement across the DNS, the hugely distributed database
system called the DNS.

So having been disabused of that, and you should note that because
there isn't just a simple technical solution to this.  And even if
there were -- and some people say, well, DNAME, DNAME solves all the
problems, I'm not sure the technical people would agree but even if
they did, DNAME is not fully deployed across the DNS, so it would take
time to deploy and to implement.

So there ain't an easy solution here.  There's a hard solution, which
is to find out what the problem is and then to address that problem.

It's probably at the registry, at the application layer.  Maybe if
we're very lucky, there will emerge a technical solution, but I
personally now think that's unlikely.

The other thing, and I said this at the beginning of the Saturday
sessions, is I think we need to be conscious to reduce the complexity
of the problem as much as we can so as to maximize the likelihood that
some IDN variants -- some IDN variants will be delegated.

The perfect is the enemy of the good.  Peter Dengate Thrush mentioned
that in relation to the resolutions that were agreed this morning.

We can search for the perfect.  We can search for solving all the
issues, and we may find that if we try and do that, that we'll never
get a solution.

Now, having said that, I must remind everybody the first thing is to
identify the issues, all the issues.  But we should be mindful of
categorizing those issues so that we can perhaps, in the next phase,
find solutions for the most important ones that affect the largest
number of Internet users.

So let's have in back of our minds that we need to reduce the
complexity, if at all possible.

So with those sort of caveats to the project and to the community, I'd
like to invite anybody who has a question or a comment to come up to
the microphones and to address the panel.

And you don't have to all rush at once.

>>CHRIS DILLON:  Hello.  It's Chris Dillon from University College,
London with a question to Dr. Govind about some Devanagari issues that
came up a couple of days ago.

And there was talk of there being problems with the Devanagari script
because several languages use that script, and there was concern that
they would be applying for identical strings and there would be major
problems with variants.

But I was just wondering if you could explain that in greater detail.

>>DR. GOVIND:   I will ask my colleague, Dr. Kulkarni, to explain
that.

>>MAHESH KULKARNI:   The Devanagari, as you rightly said, is a script
and there are a couple of languages which are schedule languages are
being represented using Devanagari.

The issue here is that the core page is the same while the languages
are being derived from that.  And we have done a study of the variants
basically.  We don't see any problem in relation with the -- basically
the variants which are coming out across the languages.

For example, if you want basically an IDN for Marathi, then
automatically for the Hindi language, it will not be a given.

So that's where we are trying to address that.  And because from the
end-user perspective, you don't have on the URL bar, basically, you
don't have a language identification system.  So Devanagari, as it
appears as a script, and whose server applies first?  Either the Hindi
people or either the Marathi people or either the rest of the
languages people, the first come, first served?  This is what exactly
we are plan.

I hope I have answered your query.

>>CHRIS DILLON:  Yes.

Now, one, just very quick, additional question.  The Punjabi has a
script, a Gurmukhi script, and is that regarded as being part of
Devanagari or is that separate?

>>MAHESH KULKARNI:   No, actually, we have got 22 Indian languages,
Shudhu languages.  There are many more.  And some of the languages
have been represented by two different scripts.

For example, if you take Punjabi, it has been written using Gurmukhi
as well as using Shahmukhi.  But Shahmukhi is like Arabic style of
writing basically.

But since it has not been currently incorporated into the Unicode, we
are not taking that.  So we will be considering only Punjabi, which is
written in Gurmukhi script.

>>CHRIS DILLON:  And you would not use the word Devanagari to refer to
--

>>MAHESH KULKARNI:   No, no.  They are totally two different scripts
basically.  We have got, for 22 Indian languages, we require something
close to around ten different scripts are there to represent all the
22 languages.

>>CHRIS DILLON:  Thank you very much, indeed.

>>DENNIS JENNINGS:   Thank you, Chris.  We have a questioner in the
middle, please.

>>ANDREW MACK:   Thank you.  Andrew Mack.  To your discussion of the
different languages that are -- the different scripts that are used
for the similar languages, we've talked in the past about the idea of
trying to bundle those together so that an applicant might be able to
apply for two or more scripts to use -- that are used to express their
language.  And I was curious about your thoughts about that in the
case much India.  And also, what would the cost be of that?  You know,
so if you can add in a second -- what's the actual cost of that?  Not
the 185 but the actual technical cost of adding a separate script?

>>DENNIS JENNINGS:   Microphone.  Dr. Govind, the microphone is off.

>>DR. GOVIND:   Ram would like to answer.

>>RAM MOHAN:   This is Ram Mohan.  I guess I didn't understand when
you said what is the technical cost of adding another script.  You
would simply instantiate -- if you were running a registry you would
simply instantiate another script.  And whatever the TLD label is, you
would simply run it that way.

Now, I don't think, so far, there has been much discussion about the
combination of multiple scripts and calling those different scripts as
variants of each other.  I don't think we've got there yet.  And I
don't think that's what Dr. Kulkarni was talking about either.

So if you look at, for example, the Devanagari example, you have
Devanagari, and as Dr. Kulkarni was saying, you have Hindi and you
have Marathi in there.

The idea is, for example, Bharat, which is India in Hindi. If you want
to get Bharat in Marathi, or if you want to get a second-level domain
name in there, that's a different problem set all together.  But if
you would like to get a domain name, whatever, dot XYZ, in Devanagari
and you want to get that same string, XYZ, in, say, Tamil, there is
not a plan or a proposal to treat them as a bundle.

Eventually, you would end up having two separate registries.  You
would end up running it through the registrars, and you would offer it
as two unique sets of strings.

So I hope that answers your question.

>>DENNIS JENNINGS:   Thank you, Ram, for that explanation.

The word bundling, I've heard, has been bandied around quite a bit.
Like all these terms, I'm not quite sure what it means.  I think we
need to define bundling, and then we need to work out whether this is
just a self-serving statement of interest or whether there really is
an issue here, and whether it belongs to our project or not.

So yet another challenge to us to define bundling and to see what it
means.

Patrik, we have a question.

>>PATRIK FÄLTSTRÖM: Patrik Fältström.  I am chair of the SSAC,
Security and Stability Advisory Committee.  I just want to make one
remark in the process we have here.  I am really, really happy to see
this project starting, let me start there, and I am also really happy
with the quality of the discussion here at this session.

I'm one of the authors of the IDN standards.  I have been fighting
with these kinds of problems myself for many, many years.

But, for example, I hear that people at this session, one of the first
time in the ten years I am working at this, I hear people really
separate the definition of script and language.  Well done.

Okay.  My remark regarding the process.  As the chair of SSAC, I just
want to make everyone understand that there are many, many individuals
that are also members of SSAC, including staff and also volunteers
that are part of this project.  But just because we have that, I do
envision in the future that this project might actually ask specific
questions to SSAC.

So I just want to point out to everyone that we don't have any
liaisons formally between SSAC and this project, but we have
individuals that participate in the project.  So when we later might
have questions that are sent over to SSAC or if it's the case that
this group, for some reason, would like to have a formal liaison,
because we at SSAC do very explicitly separate incoming and outgoing
liaison relationships with other external groups at ICANN, in that
case I am really happy to have that discussion and create those
liaisons.  Because I want these individuals to participate in these
projects to be able to use their individual skills, their individual
knowledges to participate fully in these projects and not feel they
have to be bound neither now or in a potential future discussion
inside SSAC when we try to answer some of these questions with the
work that the previous have done.

Thank you.

>>DENNIS JENNINGS:   Patrik, thank you very, very much for that and
that's very helpful, indeed, and you and I should have a discussion
off-line to flesh that out a little bit and see when that's
appropriate to set up the liaisons, if and when.

Thank you.  I appreciate that.

We have an online question, I believe.

Francisco.  Oh, Naela, you are going to take the question.  Thank you.

>>NAELA SARRAS:  Thank you, Dennis.  We have a question from Mr. Paco
Fernandez.  The question is can VeriSign now apply for new IDN gTLDs
-- for example, Russian and Hebrew -- or would they be held up by this
variant issue?

>>DENNIS JENNINGS:   And Francisco, do you want it take that question?

>>FRANCISCO ARIAS:  Sure.  I believe the simple answer is yes, they
can apply.  They are not going to be held up.  Anyone can apply for an
IDN, would be able to apply for an IDN gTLD.

What we are discussing in this project is the variants of IDNs.

>>DENNIS JENNINGS:   And if I may add to that, as I said at the
beginning, if there is a variant, you may apply for a string and state
a variant.  If you look carefully at the Applicant Guidebook, um see
that.

I think there's some more work to be done on -- in that area.

So if you have variant strings, although variant TLDs will not be
delegated until this project has produced the issues report and then a
subsequent project has looked at solutions, there is, nevertheless,
the opportunity in the first round to apply for a TLD and to name a
variant.  Just to make that clear.

Have we any other questions from the audience?

Yes, we have a questioner.  Please come up to the microphone and
identify yourself.

>> My name is Promot Pondhi (phonetic).  I come from Jalalal
(phonetic) University.  This is my first ICANN meeting.

A question I have is for the members in general.

I wonder if we want to identify the issues of variation variants.  If
you can arrive at something like consensus of the classifications --
for example, looking linguistically at the Devanagari script, I can
think of two clear sources of variants.  One I would consider
historical; that is, we have characters which stand for sounds which
are no more there.  But because, say, the language using it today --
for example, Hindi -- has relation with Sanskrit, so although the
sounds have changed, but the letters have been kept intact.

So you have examples, letters of a Gaelic ru (phonetic) kind of thing
which is not pronounced.  Shajitu (phonetic), you can write it in two
different ways.

The other has to do with the relevant orthographic level.  I think of
since Hindi, Devanagari is a syllabic script, so you have something
like a syllable represented by one grapheme, so things like Chun in
Chunden (phonetic) and Chum in Chumpeck (phonetic).  The variation is
on account of treating Chun as a phoneme, and so you have something
called a dot Bindhu, but if it is not phoneme, if it is to do with
surface representation, then you have the actual nasal consonants,
like nah, mah (phonetic), et cetera.

So the variations are on account of referring to different
orthographic levels.

So I can think in Devanagari, at least, of two different sources of
variants, orthographic level.  So these graphemes result -- variation
results because they refer to two different levels, and historical
sort of remnant sounds.

So I wonder -- I believe the languages, scripts such as Latin and
Greek may have similar problems.  I have no clear idea.  But some
consensus, at least, something like sharing of discussion, dialogues,
can lead to the consensus on types of -- sources of variants and
solutions to them.

Thank you.

>>DENNIS JENNINGS:   Thank you for that.

Yes, one of the things that we have discovered is that we need to have
lots of examples to clarify what we're talking about, and those
examples are helpful.  So there's a lot of work to be done here.

Dr. Govind or Dr. Kulkarni, do you want to take an initial response to
that question?

>>MAHESH KULKARNI:   Yeah, actually, what Dr. Pondhi (phonetic) was
referring to, variants, basically, especially coming from two
different streams.  One is the variant coming from the phonetic
variations which comes up and which we term as homophones.  And there
is another variation which comes from the -- what is called the way
one looks at the screen, which means that homo graphs.

Both aspects were considered and debated.  We did this study for
almost last four to five years, and we came to the conclusion that
homophones -- treating homophones will be a difficult task and we
might debar a lot of what is called the possible IDNs.  And we have
restricted ourselves only to the homographs.  And the latest Unicode
version, 6.0, caters to what is called the language nuances as well
the script, which is properly represented, the languages which are
properly represented using the Devanagari script.

So I don't see we should, at least for the Devanagari, we should not
treat homophones.  We are only going like homographs, as is the case
with the ASCII, like color and C-L-O-R and C-O-L-U-R, and we treat
them differently.

A similar case lies with Devanagari.  Like Hindi, are written in two
different forms.  Both are acceptable forms.  With Bindhi (phonetic),
that is with Donaswar (phonetic), and with Nahallendar (phonetic).

So these are two forms, but since they are not similar looking, we
have taken a decision to keep them separated.

So I hope that answers the question.

>>DENNIS JENNINGS:   Thank you.

Any other case study coordinator want to comment on the -- Nicholas.

>> Could I make a method lodge cam point about this.  Obviously
everything that has been said is absolutely correct.  But as I
understood the purpose of this activity, it is to identify potential
cases, to generalize over them and to see what principles might be
appealed to in future in adjudicating whether things are variants or
not.

We are not giving solutions at this stage, but the more examples we
consider, the more powerful and effective and useful in the future the
principles we arrive at will be.  And of course you will find
different principles in respect of different scripts.

And probably there will be more principles to scripts we are not
considering as well but that's for the future.

Thank you.

>>DENNIS JENNINGS:   Thank you.  Any case study coordinator want to
comment on homophones rather than homographs?  I think it's been very
well covered by Dr. Kulkarni.  So seeing no volunteer and seeing that
we have no questioner on the floor, do we have any further questions
from the -- excellent.  With five minutes to go we can clear the room
for the next session.

Thank you very much, indeed -- oh, Ram, you have a question?

>>RAM MOHAN:   Not a question, actually.  Just a comment.  For very
helpful suggestions like Dr. Pondhi said, Dennis, it would perhaps be
useful to put up on the screen or someplace where these folks can
actually continue to participate and provide input into the various
case study teams because this is a community effort and what's need
aside that kind of multiplicity of opinion.

Whether or not, you know, having two -- considering two sources is a
generalizable problem is one thing, but having that kind of input come
in is really the critical core of what this entire effort is about.

Thank you.

>>DENNIS JENNINGS:   Ram, thank you very much, indeed.  You are
absolutely right.  This -- To emphasize, this is a community project.
It's not an ICANN project.  It's a community project.  The case study
teams are community teams.  And there's an opportunity for everybody
in the community to participate.  And Francisco has brought back up on
the screen the mailing list and the Wiki.  The mailing list for
participation, the Wiki for information that will be published.  And
it's our intent, while we will be having -- the case study meetings
will be closed just for working purposes, the outputs will be
published for everybody to see.

And these are the mechanisms for people to ask questions, to raise
questions, and to contribute.

And we certainly hope, because I emphasize, these are community
projects, not ICANN projects, that people, as we have seen, will ask
questions and will contribute.

For ICANN, I would like to say that I have been -- and I think I can
say this as a consultant and not as a full staff member, that I have
been tremendously impressed both with the case study teams and the
work they're doing and with the staff, and I think this project is in
good hands.

And with that positive note, ladies and gentlemen, thank you very
much, indeed.

[ Applause ]