Some process change suggestions

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Howdy folks,

We've discussed the fee, and the elements of the application process, 
and where forward progress may terminate, and the unexpended portion of 
the fee for the portions of the process not expensed, and refunds.
I want to take a figurative step back from the draft we've been engaged 
with since the Council and Staff workshop last February in Los Angeles.
Overall I'm concerned with the choice of creating a single legal 
instrument which is incapable of distinguishing between proposals for 
unrestricted and restricted on-going operation, whether we call the 
restrictions "sponsored" or "community" or even the last-minute 
over-the-transom thing that some ideologically engaged proponents of 
"efficiency" and "deregulation" at Charles River managed to cobble 
together. The specific terms, which Susan Crawford spoke to at Cairo in 
her final address to the community as a member of the Board, and which 
David Maher has summarized for the other registry operators, are 
adequately illuminated for correction.
I'm also concerned that the choice of creating a single process which 
is, by design, incapable of distinguishing between applications intended 
to provide service to classes of under-served users of the DNS, and 
applications intended to game either the application process, or game 
existing registries providing service to users of the DNS, and worse, 
moves the savings produced by approve-capable applicants, at their cost, 
to to costs produced by reject-capable applicants, at their savings. I 
made this point during the Sunday GNSO and Staff session to Doug Brent, 
that care and cost in application preparation will, under this model, 
subsidize the lack of care and the lack of cost in application preparation.
Basically, every half of every applicant dollar spent on care is 
transferred as savings for applicants spending less on care, and the 
good subsidize the bad.
To take a specific example everyone is aware of, there will be a 
community-based application for .cym. It is not worth a nickel to "find 
out" if .cym is sufficiently similar to .com, and the Welsh linguistic 
and cultural applicants should not share the costs of developing a 
process for identifying the applications for unrestricted registries 
exploiting string gimmicks to be sufficiently similar to .com to create 
confusion and steal value.
There should be a "didn't use" box on the applicant case management 
process form, and applicants refunded or credited the estimated cost of 
having chosen not to go into that well-known box.
But more important than pay-as-you-go, which we all fully appreciate is 
very difficult for Staff to completely cost out, to replace a flat fee 
inflated for the cost of identification of economic gamers and the 
(rare) applications that may give rise to real security and stability 
concerns, which in effect is an invitation to bad actors and a barrier 
to good actors, exactly the wrong process design goal, is who, when and 
how evaluation occurs.
I've no concerns handing my applications to Chuck or David or Jeff or 
Carolyn. I have no concerns handing my applications to Tim or Paul or 
Jon or Bruce. I can't, off the top of my head, think of anyone who's 
been on the GNSO Council, and this includes people I think more wrong 
about more things more of the time than I'd like, who couldn't 
reasonably sort a pile of applications in to "yes", a "no" and a "needs 
expert or other eyes" piles.
We "insiders" can significantly assist Staff in triage. There is no 
rebuttal possible to this observation. We know more than Staff, and our 
interest is in helping Staff succeed.
No live test has been attempted, nor is a live test on plan. We know we 
have applications who's authors intend to be as uncomplicated as the 
.cat application, and these are being delayed until the process for the 
intentionally complicated applications is complete, so these 
uncomplicated applications are paying both money (above) and time (here) 
unreasonably. We can, without harm to the new gTLD process we have all 
worked so hard on for several years, test our collection of compromises 
and guesses with live applications, lower the time and dollar cost to 
careful applicants, and transform a drift towards adversarial outcomes 
and opaque process, also noted in the RyC comment, into collaborative 
outcomes and a more transparent process.
This note then is the longer, public comment version of my Sunday GNSO 
and Staff session exchange with Doug Brent and Kurt Pritz, and my longer 
hallway discussion with Doug.
In the course of writing up this comment I recalled that there was a 
proposal for pre-certification of registry operators. I didn't see the 
value in that idea when I last thought of it, but since I know, from 
experience, that any proposal using VGRS, NS, AF, CORE, and quite 
probably the resources of several current registrars, and even some 
ccTLD registries, is presumptively "safe and sane" for the common 
services models, and therefore the cost and complexity of technical 
evaluation for such applications should be close to zero. I vetted the 
.coop registry, vetting the meaningful part of the applicant's 
capabilities was just a few days of my time, and doing it repeatedly 
would improve the evaluator's best use of time.
Independent of any rule changes we make as part of reviewing Staff's 
draft, there are some broad process changes that we can make that will 
reduce the difficulty of conducting the round, and reduce the cost in 
time and cash of careful, correct applications, and free resources for 
the more difficult evaluation of applications which surprise us, some 
for good reasons, some not.
Eric Brunner-Williams
CTO, CORE