[arabic-vip] ZWNJ (earlier: [vip] Overarching principles used in Devanagari team report)

[Siavash Shahshahani]

Dear Dr. Al-Zoman,
I've seen your statements about ZWNJ and noted the wholesale changes and
deletions you are proposing for the final document. I wished you had had a
chance to participate more fully in the latest rounds of discussions that
led to this document. It is truly a tribute to Sarmad's infinite patience
and fairness as well as to  hundreds of hours of work by him and others
that this document is seeing the light of the day. Certainly some
fine-tuning and corrections may be in order, but such drastic deletions and
changes as you're proposing is tantamount to rejecting this very
commendable effort. I sincerely believe that a little discussion can
alleviate your present misgivings, and you will join the rest of us in
celebrating the end of the first phase of this endeavor. I'll make just two
comments to address your major concerns:
1. This document is NOT solely concerned with putting TLDs in the root. It
is true that much of the discussion on the structure of variants
concentrated on the top level, but that's because the top level domain
structure is the part directly under ICANN's purview and there is a new
gTLD delegation process coming up for which ICANN must have clear and
immediate guidelines. On the other hand, gTLD delegation process is more
than just putting a string in the root. There are, e.g., legal aspects to
the gTLD process such as dispute resolution for which ICANN has shouldered
responsibility. It is imperative that the dispute resolution regime under
which new gTLDs are to operate be defined beforehand; you cannot dismiss
that effort by fiat. You may believe, as I do, that very little
modification is needed in the present UDRP to make it operable for IDN
gTLDs, but this has to be demonstrated through a thorough examination of
possible problems that variants and other aspects of IDN may engender. Same
goes for several other topics you had expressed a  wish to eliminate. 
2. I believe the fear experienced by some about the use of ZWNJ is highly
illusory. Suppose you see a string on a bill-board or in the airport
containing a ZWNJ. It may confuse you as an Arab because you don't use ZWNJ
and we've all been taught that empty space is not allowed in a domain
string. But how does this threaten the security and stability of the
Internet? You will certainly not be led to reaching a wrong address,
because typing the empty space will not get you anywhere. And why should
you bother? You won't probably need that address anyway. If the company
with that domain address wants to reach the Arab audience, it will take
care of adopting a domain reachable by Arab keyboards. Please note that we
eliminated the confusion that may arise in the use of ZWNJ in conjunction
with the three characters U+0637-8 and U+069F. You don't really want to
disenfranchise some 300 million people who use the Arabic script WITH ZWNJ?
I think Behnam's example of the confusion between < بیا > and < یبا >
should be more threatening. Both are perfectly legitimate strings on any
Arabic-script keyboard. 
I hope this little friendly explanation will help.
Best regards,
Siavash
 
On Sun, 25 Sep 2011 10:04:46 +0000, Abdulaziz Al-Zoman
<azoman at citc.gov.sa>
wrote:
> Dear Dr. Sarmad and all,
> 
> You stated that " Again, the choices being proposed accommodate the
needs
> of all language communities, where Arabic language community may decide
not
> to activate the variants with ZWNJ and the communities of other
languages
> may decide to activate them. "
> 
> It is foreseen that new gTLD registries would not be attached to single
> community. There will be multiple international companies that will
provide
> Arabic Script TLDs. They will serve the whole script communities.
> 
> Therefore, if ZWNJ is allowed in a TLD then it is not the choice of the
> language community to activate or not activate the variants with ZWNJ.
> 
> 
> -----------------------------
> عبدالعزيز بن حمد الزومان
> Abdulaziz H. Al-Zoman
> 
> From: sarmad.hussain at kics.edu.pk [mailto:sarmad.hussain at kics.edu.pk] On
> Behalf Of Dr.Sarmad Hussain
> Sent: Sunday, September 25, 2011 11:12 AM
> To: Abdulaziz Al-Zoman
> Cc: arabic-vip at icann.org
> Subject: Re: [arabic-vip] ZWNJ (earlier: [vip] Overarching principles
used
> in Devanagari team report)
> 
> Dear Dr. Al-Zoman and All,
> 
> What you would suggest would be an option 4.  However, option 3 is
> intended as it is articulated (different from option 4).
> 
> Options 1-3 are suggested to cover the breadth of needs (including those
> for Farsi, Kurdish, Urdu and many other languages being used online
using
> the Arabic scrip).
> 
> ZWNJ is already prohibited in contexts it is invisible.  It is
indirectly
> visible in all the contexts being suggested due to shaping alternation.
> 
> Again, the choices being proposed accommodate the needs of all language
> communities, where Arabic language community may decide not to activate
the
> variants with ZWNJ and the communities of other languages may decide to
> activate them.
> 
> As per the activation, in our F2F meeting, we discussed that the variant
> label set could be pre-determined through the Arabic script TLD label
> generation policy but which subset of these variants is activated is
> requested by the applicant (others, please correct me if this is not the
> correct understanding).
> 
> 
> 
> regards,
> Sarmad
> 
> 
> On Sun, Sep 25, 2011 at 12:28 AM, Abdulaziz Al-Zoman
> <azoman at citc.gov.sa<mailto:azoman at citc.gov.sa>> wrote:
> Thanks for the clarification …
> 
> What I'm with is the option that:
>         What you type is what you see.
> 
> When a user see a label in an airport, for example written in Arabic
> script,  he/she would like to type it as it appears in the add. ZWNJ/ZWJ
> are transparent to the users they cannot be seen and cannot be typed by
the
> (Arab users). It looks like a space (which is not permitted in domain
> names). Hence, if the add displays a domain with a TLD which includes
ZWNJ,
> (Arab) users will get confused and will not be able to reach that domain
> name.
> 
> So, the option 3 should not include the ZWNJ in the fundamental TLD nor
in
> the activated/delegated variants of the TLD.
> 
> -----------------------------
> عبدالعزيز بن حمد الزومان
> Abdulaziz H. Al-Zoman
> 
> From: sarmad.hussain at kics.edu.pk<mailto:sarmad.hussain at kics.edu.pk>
> [mailto:sarmad.hussain at kics.edu.pk<mailto:sarmad.hussain at kics.edu.pk>]
On
> Behalf Of Dr.Sarmad Hussain
> 
> Sent: Sunday, September 25, 2011 9:40 AM
> To: Abdulaziz Al-Zoman
> Cc: arabic-vip at icann.org<mailto:arabic-vip at icann.org>
> 
> Subject: Re: [arabic-vip] ZWNJ (earlier: [vip] Overarching principles
used
> in Devanagari team report)
> 
> Dear Dr. Al-Zoman and All,
> 
> Just to further clarify my own email (not responding to Dr. Al-zoman's
> mail below), please note that even in option 3 it is being suggested
that
> ZWNJ can be allowed in the variant to fundamental label (up to the user
to
> activate or reserve it).  Just wanting to make that explicit.
> 
> regards,
> Sarmad
> 
> 
> 
> 
> On Sat, Sep 24, 2011 at 11:25 PM, Abdulaziz Al-Zoman
> <azoman at citc.gov.sa<mailto:azoman at citc.gov.sa>> wrote:
> 1 and 2 is not valid options for a normal Arab user at all … when he
sees
> a label which includes ZWNJ … he/she will think that is a space … the
other
> form (without ZWNJ) is NOT similar to the original one with ZWNJ … hence
> he/she cannot reach that domain name
> 
> Hence, there is only one option, unfortunately, which is 3.
> 
> -----------------------------
> عبدالعزيز بن حمد الزومان
> Abdulaziz H. Al-Zoman
> 
> From: arabic-vip-bounces at icann.org<mailto:arabic-vip-bounces at icann.org>
>
[mailto:arabic-vip-bounces at icann.org<mailto:arabic-vip-bounces at icann.org>]
> On Behalf Of Dr.Sarmad Hussain
> Sent: Sunday, September 25, 2011 9:11 AM
> 
> To: arabic-vip at icann.org<mailto:arabic-vip at icann.org>
> Subject: Re: [arabic-vip] ZWNJ (earlier: [vip] Overarching principles
used
> in Devanagari team report)
> 
> Dear All,
> 
> Thank you for your condolences privately and on this list.
> 
> 
> We have seen the opinions on ZWNJ on this list both in support of and
> against having it enabled for TLDs.
> 
> Firstly, let me reiterate that these arguments clearly brings it out as
an
> issue (will try to summarize all arguments in its favor and against in
our
> final document, but not doing it here as (even if we may not agree) all
of
> us understand the arguments being made in each case).
> 
> However, I would want to go beyond and suggest a possible
recommendation,
> which (to me) can actually address both sides of argument
simultaneously:
> 
> 1. If we recommend that ZWNJ is allowed, however the string with it is
> considered a variant of the string without it, that addresses KB,
> confusability and security issues (but gives the users the choice and
> flexibility based on their language)
> 
> 2. If we want to be more conservative, we can suggest that if ZWNJ is
> allocated, then the variant without it must also be allocated
> 
> 3.  If we wan to be even more conservative, we can also suggest that the
> label with ZWNJ cannot be a fundamental label
> 
> 
> Could this address the issue?  If yes, should we stop at 1? 2? or must
> also have 3?
> 
> You are kindly requested to consider both sides of the arguments
> sympathetically as you respond.
> 
> 
> regards,
> Sarmad
> 
> 
> 
> 
> On Sat, Sep 24, 2011 at 9:58 PM, Abdulaziz Al-Zoman
> <azoman at citc.gov.sa<mailto:azoman at citc.gov.sa>> wrote:
> Dear all,
> 
> I do agree with what Andrew has said:".. the mere
> fact that someone sometimes uses a string as part of their language is
> in no way an argument that such a string should be permitted as a
> label at any part of the DNS, never mind in the root"
> Hence, digits, space, and ZWNJ are not allowed just because they will
> cause more harms than benefits. Here we are dealing with the TLDs at the
> root level hence we should give security, stability and usability the
> highest priority.
> 
> Therefore, I did recommend (previously) not to support the space in the
> TLD and now I strongly with NOT to support ZWNJ/ZWJ at the root TLD
level.
> 
> Just a note:
> The ZWNJ/ZWJ as a concept is not know at all by the Arabic speaking
> community even not be the ARAB IT Specialists:
> - They (the Arab users) do not know their (i.e., ZWNJ/ZWJ) behavior.
> - They never heard about them.
> - They never type them.
> - It is not in the keyboard.
> - They do not know how to type them.
> - and most importantly ... they cannot see them.
> 
> 
> -----------------------------
> عبدالعزيز بن حمد الزومان
> Abdulaziz H. Al-Zoman
> 
> -----Original Message-----
> From: arabic-vip-bounces at icann.org<mailto:arabic-vip-bounces at icann.org>
>
[mailto:arabic-vip-bounces at icann.org<mailto:arabic-vip-bounces at icann.org>]
> On Behalf Of Andrew Sullivan
> Sent: Sunday, September 25, 2011 4:07 AM
> To: arabic-vip at icann.org<mailto:arabic-vip at icann.org>
> Subject: Re: [arabic-vip] ZWNJ (earlier: [vip] Overarching principles
used
> in Devanagari team report)
> On Sat, Sep 24, 2011 at 11:09:41AM +0000, Raed Al-Fayez wrote:
> 
>> So I think if the group is going to support ZWNJ and/or ZWJ in
>> TLDs(because it may be needed by other languages) then we should
>> also ask to support/reconsider the support of the space in TLDs and
>> even more in all Internet Standards (IDNA RFC, DOMAIN RFC
>> ..etc). Also if the group is going to support them in the TLD then
>> we should add a warning that they may cause a risk and ICANN should
>> study this topic carefully.
> 
> Just to be clear: what you are requesting there is that the team
> recommending throwing out restrictions in RFC 952 (and consequently,
> depending on your reading, RFC 1035), and also restrictions built into
> IDNA2008 (RFC 5890-5894).  This amounts to a requirement that every
> machine on the entire Internet needs to be upgraded before the change
> could take place.  Are you sure that's something you want to ask for?
> 
>> Personally I think enabling the numbers in TLD is also important
>> because we are not sure that other language may use them in their
>> writing system (I hear that Jawi language do so).
> 
> I have suggested before, but I will suggest here again, that the mere
> fact that someone sometimes uses a string as part of their language is
> in no way an argument that such a string should be permitted as a
> label at any part of the DNS, never mind in the root.  Irish names in
> English and French all over the place both use the apostophe, U+0027.
> That is not an argument that we should start allowing that character
> in to DNS labels (even though, in a strict sense, it is already a
> perfectly legal character).  The reason digits aren't permitted at the
> top level is simple: they're potentially confused by end system
> software as IP addresses.  (Dotted quad is not the only way to
> represent IP addresses, note.)
> 
> Best regards,
> 
> A
> 
> --
> Andrew Sullivan
> ajs at anvilwalrusden.com<mailto:ajs at anvilwalrusden.com>
>
-----------------------------------------------------------------------------------
> Disclaimer:
> This message and its attachment, if any, are confidential and may
contain
> legally
> privileged information. If you are not the intended recipient, please
> contact the
> sender immediately and delete this message and its attachment, if any,
> from your
> system. You should not copy this message or disclose its contents to any
> other
> person or use it for any purpose. Statements and opinions expressed in
> this e-mail
> are those of the sender, and do not necessarily reflect those of the
> Communications
> and Information Technology Commission (CITC). CITC accepts no liability
> for damage
> caused by this email.