Re: [ALAC] Fwd: ICANN News Alert -- New gTLD Batching Announcement
At 07/06/2012 11:22 AM, Evan Leibovitch wrote:
I really don't give a damn about the method of random selection of who gets in what batch, with one exception stated below. There are many less-gamable ways to do the selection, based on data that has already been collected.
Here's one super simple example that most people on this list could implement (using open source software for maximum transparency): 1) make a checksum of each received application document that is in the pool (and publish the list of checksums) 2) create a random checksum (seeded by the moment Rod presses "Enter" at a public event staged for the purpose) 3) Select the 500 applications whose checksums are numerically closest to the random one.
I believe that sine a "random number" is involved in the, it would be considered a lottery under California law and is thus not allowed.
The above example is far from perfect and anyone taking more than a minute or two could come up with something better, but even this is preferable to ICANN playing William Tell. It's transparent, and it eliminates the pool.comoffering since the batch choices are made based on pre-existing data that did not anticipate this process, and no further action on the applicant's part is necessary.
Still.... to me this is a fight between applicants, and IMO there is little public interest in their competitive issues. My only concern is about the appearance that ICANN has chosen a process that service providers are claiming is game-able, and that appearance is damaging to ICANN's reputation.
Agreed.
I agree with Olivier that simply because orgs are promising ways to game the system doesn't mean they can. Even a psychic can be correct on occasion. But it's the *appearance* that the system is so easily game-able that sends such a horrid message outside the ICANN bubble.
So what is there for ALAC to say?
ICANN in its (lack of) wisdom refused to consider the concept of categorizing applications before the fact, so the option of determining which subsets of applications are in the public interest is not available without starting the evaluation process. All we have is the "community" designation and IDNs. And ICANN has already said that applications requesting support won't be done in the first batch.
Personally, I would limit any comment of ALAC's to 1) Demand that community and IDN applications be given advanced priority as a matter of public benefit 2) Repeat our warning of the bias against developing economies evidenced by the relegation of applications requesting support 3) Note the damage to ICANN's reputation created by the choice of a selection method that can appear to be gamed
My personal feeling is that there are plenty of people criticizing this and given our workload and the timing, this is one we can afford to skip. Alan
- Evan
On 7 June 2012 06:51, Olivier MJ Crepin-Leblond <ocl@gih.com> wrote:
Hello all,
please find below, more information about the Digital Archery which will be used to batch new gTLD applications into batches of probably 500 applications at a time. With a possible 2000 applications to process, some applications might only be processed much later than others and this "game" of Digital Archery will set applicants apart. Already several firms run by close ICANN insiders are proposing a paying service to "win" the Digital Archery using their automation service which they claim will ensure their client's applications are considered first. This means more sources of profit, but this also means that applications which were made possible by thanks to all of the work the community did to have Applicant Support, might be discriminated against since they will not be able to afford the automated digital archery winning service.
Also, looking closely at the terms and conditions of "digital archery", (see video on http://newgtlds.icann.org/en/announcements-and-media/video/batching-demo at Time 2:37 mark) , applicants using the Digital Archery system agree legally that this is a contest of skill rather than a contest of chance. Clearly, with the Internet's differing delays depending on where you are located, and since digital archery system will be accurate to 100th second (as seen on the demo), it is *impossible* for anyone to display any kind of skill except by being physically at the data center itself. I remind you that delays on the Internet range from 3ms to 700ms or more, depending on where you are, since any satellite link will automatically induce between 300 to 700ms -- it is a physical propagation time needed. And these delays are jittery too.
Is Digital Archery therefore allegedly a game of chance, not controlled by any kind of legal structure associated with games of chance, with open opportunities for cheating?
There have been calls to the Board by some members in the ccNSO for Digital Archery *not* to be used and an alternative process to be followed, which may also involve *not* having any batching altogether. Others in the GNSO Council have also written to the Board about this, with similar findings or other alternative propositions. Should we support those comments?
Alas, the Board correspondence page has its last entry on 21 May 2012 and the correspondence which I have been CC'ed into was sent on June 4th and on June 6th but I have asked if those have already been published.
With Digital Archery starting on June 8th, should the ALAC issue a statement on this ASAP?
I'd be interested in your points of view.
Kind regards,
Olivier (the above message includes some of my own points of view)
-------- Message original -------- Sujet: ICANN News Alert -- New gTLD Batching Announcement Date : Wed, 06 Jun 2012 22:35:29 -0400 De : ICANN News Alert <communications@icann.org> Pour : <ocl@gih.com>
ICANN <http://www.icann.org/>
News Alert
http://www.icann.org/en/news/announcements/announcement-06jun12-en.htm
------------------------------------------------------------------------
New gTLD Batching Announcement
6 June 2012
Given the large number of new generic top-level domain (gTLD) applications, we will divide and evaluate them in batches. The batching system is targeted to open at 00:01 UTC on 8 June 2012, and will close at 23:59 UTC on 28 June 2012. The target date for posting the batching order is 11 July 2012.
The batching process will be used to determine which applications will be processed in the first batch, the second batch and so on. It will be done by: assignment of a timestamp, and the formation of batches.
/Timestamp assignments/ will be done using the TLD Application System (TAS). All applicants must use their TAS credentials to log in, read and accept the batching rules, indicate their batching preference, and select their target date and time. Once these steps are completed applicants should log back into TAS to hit the target time and generate a secondary timestamp. Users will have access to a testing feature to gauge the secondary timestamp system's response time.
/Batching formation/ considers an applicant's: (1) batching preference, (2) geographic region and secondary timestamp; and (3) contention among identical and "similar" applications.
(1) Applicants stating a preference for "opting-out" will be placed last.
(2) Geographic diversity is important in bringing more competition and choice into the domain name market. Applicants who opted in will be ranked within their geographic region (Africa, Asia-Pacific, Europe, Latin America/Caribbean and North America) by their secondary timestamp score. Then applications will be selected from each ICANN region using a "round robin" approach. This approach selects the best timestamp score from each region, one region at a time, on a rotating basis. If a region runs out of opted-in applications, the "round robin" continues across the remaining regions. This process continues until the batch is formed, with the opted-out applications last.
(3) ICANN will then make preliminary determinations of contention sets based upon exact match. All applications in a single contention set are placed into the batch where the earliest application in the contention set is placed. Once the string similarity panel establishes complete contention sets, "similar" strings might be reassigned to an earlier batch. No applications will be demoted as a result of the promotion of others. This could result in a batch larger than 500.
ICANN has taken care to provide a secure and stable platform for the batching system. Users will connect to the Citrix XenApp high-availability cluster and will then log into the batching system. Applicants will be required to agree to a set of Batching Rules, including an agreement that "ICANN reserves the right to delay an application to the last batch or to reject an application entirely if ICANN reasonably determines that the applicant abused the batching system or intentionally interfered with the performance of the system or any other applicant's use of the system."
Along with this announcement ICANN is posting several additional resources to inform applicants about the batching process. These include a: set of frequently asked questions (FAQs), video demonstration, user guide, batching details and rules, and a batching basics fact sheet, which all can be found on the Batching information webpage at http://newgtlds.icann.org/en/applicants/tas/batching/. Information on security, infrastructure, and operations is also available in these materials.
This message was sent to ocl@gih.com from:
ICANN | 4676 Admiralty Way Suite 330 | Marina del Rey, CA 90292-6601
Email Marketing by iContact - Try It Free! <http://www.icontact.com/a.pl/144186>
<
http://app.icontact.com/icp/mmail-mprofile.pl?r=9829257&l=6333&s=YVM2&m=8845...
_______________________________________________ ALAC mailing list ALAC@atlarge-lists.icann.org https://atlarge-lists.icann.org/mailman/listinfo/alac
At-Large Online: http://www.atlarge.icann.org ALAC Working Wiki:
https://community.icann.org/display/atlarge/At-Large+Advisory+Committee+(ALA...)
-- Evan Leibovitch Toronto Canada
Em: evan at telly dot org Sk: evanleibovitch Tw: el56 _______________________________________________ ALAC mailing list ALAC@atlarge-lists.icann.org https://atlarge-lists.icann.org/mailman/listinfo/alac
At-Large Online: http://www.atlarge.icann.org ALAC Working Wiki: https://community.icann.org/display/atlarge/At-Large+Advisory+Committee+(ALA...)
On 7 Jun 2012, at 12:08, Alan Greenberg wrote:
At 07/06/2012 11:22 AM, Evan Leibovitch wrote:
I really don't give a damn about the method of random selection of who gets in what batch, with one exception stated below. There are many less-gamable ways to do the selection, based on data that has already been collected.
Here's one super simple example that most people on this list could implement (using open source software for maximum transparency): 1) make a checksum of each received application document that is in the pool (and publish the list of checksums) 2) create a random checksum (seeded by the moment Rod presses "Enter" at a public event staged for the purpose) 3) Select the 500 applications whose checksums are numerically closest to the random one.
I believe that sine a "random number" is involved in the, it would be considered a lottery under California law and is thus not allowed.
Or at least that is the ICANN Legal staff's contention. Has anyone ever seen the argument that proves this to be the case?
...
I agree with Olivier that simply because orgs are promising ways to game the system doesn't mean they can. Even a psychic can be correct on occasion. But it's the *appearance* that the system is so easily game-able that sends such a horrid message outside the ICANN bubble.
Well those offering the service are clever. Like the house in a casino they win because statistics are on their side. Half of their clients would pay them even if they did nothing. I hope the digital archery services are bonded, otherwise anyone who trust them with their TAS account and password is a daredevil - not to mention the liability on the DA service provider itself. I would not be surprised if the DA episode buys a few more lawyers their dacha.
So what is there for ALAC to say?
ICANN in its (lack of) wisdom refused to consider the concept of categorizing applications before the fact, so the option of determining which subsets of applications are in the public interest is not available without starting the evaluation process. All we have is the "community" designation and IDNs. And ICANN has already said that applications requesting support won't be done in the first batch.
Personally, I would limit any comment of ALAC's to 1) Demand that community and IDN applications be given advanced priority as a matter of public benefit 2) Repeat our warning of the bias against developing economies evidenced by the relegation of applications requesting support 3) Note the damage to ICANN's reputation created by the choice of a selection method that can appear to be gamed
I find this is interesting. The only problem with using community is that this is just a claim some applications make. Unless there is a Community Priority Evaluation, no one actually determines the truth of the community claim. Or is the suggestion that now that the applications are closed and this can't be gamed since no one can now decide to become a community just for DA's sake, ICANN should just take the applicant's word for it. Another way to achieve a similar effect, which requires less trust, is to just push all the .brands to the back of the bus. Especially since they are reputed to only be applying defensively, how much can it matter. As for the bias against developing countries, that is something we all should repeat loud and often. while we all worked hard on the Applicant Support Program, the paucity of outreach on both the new gTLD program and on the Applicant support program means that the results will be far worse than I had hoped they would be. On reputation, when an organization is besieged by groups both internal and external who wish to destroy the organization, there is very little one can do about reputation. avri
At 08/06/2012 12:51 AM, Avri Doria wrote:
On 7 Jun 2012, at 12:08, Alan Greenberg wrote:
I believe that sine a "random number" is involved in the, it would be considered a lottery under California law and is thus not allowed.
Or at least that is the ICANN Legal staff's contention.
Yup. But I suspect that an outright random selection is pretty much likely to be considered a game of chance. Digital Archery, however is more interesting. If is is not game-able (as ICANN contends), then I don't see how it differs from a random lottery. If it is game-able, then we have an unfair situation. Either way it does not seem to be an appropriate solution. Alan
DA is also based on something random -- that is, whatever reference moment the selected timestamp is supposed to be closest to. The main difference between my checksum method and DA is that the applicant's "entry" -- in the form of the checksums of their applications -- already exists and can't possibly be gamed after the fact. I agree with Avri, that if the checksum method is susceptible to challenge as an illegal lottery of chance then so is DA. OTOH, there are a number of non-random methods to create the reference point for my alternate scheme. One of them could be to collect the first page of every application, make a checksum of THAT resulting (1900 page) document, and use that checksum as the reference against which to measure the others. - Evan On 8 June 2012 01:40, Alan Greenberg <alan.greenberg@mcgill.ca> wrote:
At 08/06/2012 12:51 AM, Avri Doria wrote:
On 7 Jun 2012, at 12:08, Alan Greenberg wrote:
I believe that sine a "random number" is involved in the, it would be considered a lottery under California law and is thus not allowed.
Or at least that is the ICANN Legal staff's contention.
Yup. But I suspect that an outright random selection is pretty much likely to be considered a game of chance. Digital Archery, however is more interesting. If is is not game-able (as ICANN contends), then I don't see how it differs from a random lottery. If it is game-able, then we have an unfair situation. Either way it does not seem to be an appropriate solution.
Alan
_______________________________________________ ALAC mailing list ALAC@atlarge-lists.icann.org https://atlarge-lists.icann.org/mailman/listinfo/alac
At-Large Online: http://www.atlarge.icann.org ALAC Working Wiki: https://community.icann.org/display/atlarge/At-Large+Advisory+Committee+(ALA...)
-- Evan Leibovitch Toronto Canada Em: evan at telly dot org Sk: evanleibovitch Tw: el56
participants (3)
-
Alan Greenberg -
Avri Doria -
Evan Leibovitch