Second Level Domains Policy Review Submission

From:		David Farrar
Received:	23 October 2003


1.
Q:  Are the twelve second level domains still appropriate for their communities 
of interest? 
A: Broadly speaking, yes they are.  As with gTLDs there is not always a lot of 
difference seen as between net.nz, org.nz and co.nz with many organisations 
registering a co.nz name and also many net.nz registrations which do not meet 
the official community of interest.  It is worth noting that more and more 
organisations register a name in all three major 2LDs as they are all seen as 
the three main “open” ones.


2.
Q: Should any of the open 2LDs be moderated? (.ac, .co, .gen, .geek, .maori, 
.net, .org, .school)
A: yes
comments: I have advocated for many years that school.nz and ac.nz should be 
moderated.  In fact it was the failure of the then ISOCNZ to continue the 
moderation of these domains which motivated me to get involved with DNS 
issues.  

If InternetNZ actively seeks out moderators within the Government education 
sector, I am confident one can be located.  

I do not support any retrospective moderation where current registrants 
can close their names (this should never happen due to a policy change), 
so existing registrations would be grand fathered in, but future registrations 
or change of registrants in school.nz and ac.nz would require moderator 
approval.


3.
Q: Is there any need for the rationalisation of the current 2LDs? 
A: Yes.  

The cri.nz 2LD has only 27 registrations, iwi.nz 45 and mil.nz 15 and 
all could be considered for rationalisation.  In the poll done for the 
DNC the prompted recognition rates were only 9%, 16% and 18% respectively.  
Unprompted would be even lower.

I propose that while cri.nz and mil.nz were created to recognise their 
roles in the early days of the Internet, they are not needed as second 
level domains as they could be located under govt.nz as they are both 
part of the Government.  The govt.nz moderator could either place them 
at the third level (agresearch.govt.nz) or at the fourth level 
(agreseach.cri.govt.nz) in their discretion.

Holistically a case can also be made that iwi.nz would be a better fit as 
a 3LD under Maori.nz – ie ngaitahu.iwi.maori.nz.  However I do note that 
the iwi.nz domain has been growing significantly (doubled in 18 months) 
unlike cri.nz and mil.nz which have been shrinking.  Closing off iwi.nz 
to new registrations could create issues with Iwi yet to register a domain 
name, so I do not propose closing the 2LD, but suggest InternetNZ do check 
with the moderator of iwi.nz if there is any desire to rationalise under 
maori.nz.

As above I do not support any current registrant automatically losing 
their domain name due to a policy change.  This is a very bad thing to 
be avoided at all costs.  So all that should happen is cri.nz and mil.nz 
are closed off to future registrations and the 2LDs only be deleted if 
and when empty.


    	3.1
	Q: If so, what process should be used for the removal or closing of 
	existing 2LDs?:
	A: A 2LD should be closed only, not removed unless all current 
	registrants voluntarily re-register elsewhere.

	Closure of a 2LD should mean that no new names can be registered and 
	that an existing name can not be transferred to a new registrant.  
	The only exception or clarification to this would be a change of legal 
	name or merger where the entity effectively stays the same.

	Removal of a 2LD should happen automatically once it has been closed 
	and after the last domain name is cancelled voluntarily.


	3.2
       	Q: Should any such process be included in the 2LD policy?:
	A: No.  Closure or deletion of a 2LD should be such a rare event 
	that each one has to be taken as an individual case.  This is not 
	something one should be doing all the same.  Any closure or deletion 
	should specifically be consulted on, as per this review.

 
4.
Q: What criteria (types of criteria and specific examples) should be applied 
to the creation of new 2LDs?
A: Firstly I think the criteria should be different for moderated and 
unmoderated 2LDs.  For unmoderated 2LDs, the process should be market 
driven with the key question being – will a reasonable number of registrants 
register names in this 2LD.

The geek.nz 2LD with almost 500 names showed that it was a viable 2LD.  
However setting up bananas.nz as a 2LD for those interested in bananas 
would prob only get less than 5 registrations.  So demonstrating a likely 
level of usage is a key criteria.  

The other key issue for unmoderated is minimum ambiguity.  For example one 
could set up com.nz and within months have 40,000 registrations in it.  
However I would suggest 99% will be duplicates of the co.nz registration.  
The new gTLDs set up by ICANN have had this problem with many such as .biz 
and .info having a very small percentage of original registrations.  Ben 
Edelman at http://cyber.law.harvard.edu/tlds/001/ reveals that many .biz 
names are duplicates of the .com name and held by the same registrant.

For moderated 2LDs, the focus should not be so much on how many registrations 
there might be, but how well defined the proposed community of interest is, 
how fair the moderation criteria are, and the suitability of the proposed 
moderator.  I thought bank.nz was a perfect application as the community of 
interest is in fact defined by statute, the moderation criteria was unequivocal 
– any bank registered with the Reserve Bank, and the moderator being a logical 
choice of the industry lobby group.

A proposal which might fail the criteria is one from myself proposing we 
set up farrar.nz as a moderated 2LD with myself as moderator.  A group of 
people having the same surname is not as commonly recognised as a community 
as say banks, lawyers, Maori etc.  The moderation criteria may be fair – 
surname must be Farrar, but there would not be proof that one individual 
would represent all the diverse Farrars out there.

Another issue for moderated 2LDs is to be careful that the name of the 
2LD and the scope match.  Bank.nz for banks is fine, law.nz for lawyers 
is arguably also fine but having cars.nz restricted to car dealers would 
not be, because there are also car enthusiasts, car manufacturers, car 
model clubs etc.  

Finally it would be sensible to have in the policy that applications for 
2LDs which are deemed offensive by InternetNZ may be declined.  I am 
actually the author of the policy to get rid of the ban on offensive 
names at the unmoderated third level, but at the second level which is 
tightly controlled is a different issue.  One doesn’t want to even have 
a public submission period on whether the (for example) rapistsrulez.nz 
2LD be created.  


5.
Q: Is the current process for creating new 2LDs too simple, too complex 
or about right? 
A: too complex
comments: The elements of discussion, straw poll, submissions, interim 
decision, submissions on interim decision and final decision all should 
remain, but they do not need to be consecutive and many can run 
concurrently.

I would propose one integrated period for consultation where people 
can discuss the proposal, make a formal submission on it and/or vote 
in a straw poll.  This would probably be over a time period of 4 - 6 
weeks rather than the current 15 - 24 weeks.

Then one would have an interim decision, a set period of feedback on 
the interim decision - propose two to three weeks and then a final decision.


6.
Q:  How much, if anything, should an application for the creation of a 
new 2LD cost? 
A: The current cost of $1,000 seems reasonable.  It is high enough to 
discourage flippant proposals but low enough not to be a barrier to 
genuine proposals.  The notion of full cost recovery is not possible 
as this would increase the cost to well over $5,000.


7.
Q: Who, if anyone, should be able to block the creation of new 2LDs? 
If so, why and how? 
A: Firstly the Council of InternetNZ should remain the decision making 
authority on applications.  However it should receive a recommendation 
from NZOC on any proposed creation.  Council and NZOC should of course 
only judge on how well an application fulfils policy, not what 
individual’s personal preferences are.

In the case of a moderated 2LD, any significant section community of 
interest affected by it should be able to block it.  This does not mean 
bank customers could block bank.nz but would allow nurses to black the NZ 
Medical Assn from gaining health.nz, but they would be okay with doctors.nz.

I do not support having the straw poll as a sudden death hurdle for 
applications.  I think it is important to retain the poll, but the results 
of it should form part of the feedback on the proposal, rather than be an 
automatic block if an arbitrary number or percentage is not gained.  


8.
Q: Should there be different criteria or processes for the creation of 
moderated as opposed to open Second Level Domains? 
A: Yes – detailed above for criteria.  The process should be the same though.


9.
Q: Is the level of public input in the process adequate; excessive; or 
inadequate?
A: adequate
comments: It is adequate.  It just goes on for too long.


10.
Q: Are the methods currently used to seek public input (submissions; 
discussion period; straw poll) the best available, or would you suggest 
other/additional methods? 
A: As detailed above, the process I would propose is:

a)	Proposal received by DNC with $1,000 cheque and asap forwarded 
	to NZOC and then Council with recommendation on whether proposal 
	has sufficient information to proceed etc.
b)	A initial consultation period of 4 – 6 weeks commences which 
	allows formal submissions, list discussion and around halfway 
	through a straw poll.  I would not have straw poll at the end as 
	there may be useful discussion or submissions on the results of 
	the straw poll
c)	NZOC considers feedback (including straw poll) and summary provided 
	by DNC and recommends to Council within 14 days
d)	Council makes interim decision within 14 days
e)	Submission and discussion on interim decision for 2 to 3 weeks
f)	Council makes final decision at next face to face meeting


11.
Q: Do you think direct public registrations should be accepted at the 
second level?  
A: no

	11.1
	Q: Why?  What are the advantages or disadvantages of such an approach?
	A: If one was designing the .nz TLD from scratch I would probably 
	advocate for direct registrations at the 2LD, as many other ccTLDs 
	do this successfully, it avoids having to register in multiple 2LDS, 
	and around 85% of registrations are in co.nz anyway

	But the impact on current registrants and registrars by changing at 
	this stage would be massive, and IMO the costs would outweigh the 
	benefits.

	Many registrants would feel it necessary to register their name at 
	the second level and face significant rebranding costs.  There would 
	be considerable confusion about the difference between kiwiblog.co.nz 
	(registered under DNC policies and protected) and kiwi.blog.nz which 
	would be a private registry set up by blog.nz registrant) with different 
	registrant rights.

	There would be inevitable disputes over newly available names and I 
	have no doubt court action to test any policy put in place.

	The DNC Office would need a massive increase in staff to cope with the 
	education and resolution issues around any change.

	So while allowing 2LD registrations does appeal to me, I think we are 
	far too far down the track to change policy in a way which will be 
	beneficial to registrants and not cause confusion, costs and disputes.


12.
Q: If direct registrations are allowed at the second level, what kinds of 
transition arrangements would need to be put in place for existing third level 
registrants? 
A: Existing 3ld registrants should be given first option on registering their 
name at the 2ld level.


13.
Q: What is the best way to deal with any conflicts between existing third level 
registrants if the second level is opened? 
A: I would advocate the .ca model where a 2LD name can be allocated only if there 
is no 3LD registration of it, the same registrant holds all 3LD versions of that 
name, or the different 3LD registrants agree between themselves which one of them 
can register it (would probably involve payments between parties)

The inherited .nz data is not of high enough quality and auditability to allocate 
on the basis of the oldest 3LD registration.  The suggestion it be done on the 
basis of whom has the most 3LD versions of a name holds little merit IMO.


14.
Q: Should registration fees be different (higher/lower) for Second Level domain 
registrations if they are opened directly to the public?
A: No – the registry fees should be the same for all registrations.  KISS.


15.
Q: Are there other options around the future of the Second Level that the 
working group should consider? 
A: No - the paper seems to cover them well.  


16.
Q: Are there any other issues you think the Working Group needs to consider 
in the course of the review? 
A: No, but the next step should be either a proposed new policy, or a small 
number of discrete varying proposals which stakeholders can give feedback on.