criticism for the concept overall
My comment responds to the invitation to "offer support or criticism for the concept overall", generally criticism. As a matter of method, if a resource is limited, then the "data" that there are one more applicant than that limit will accommodate, even much more than one more applicant, for the resource than the limit, convey zero additional information than the data that there are as many applicants as that limit. Similarly, if the data is that there are fewer applicants for the resource than the limit, "how many fewer" is not information, it is decoration. If, as has been suggested broadly in the technical community, the rate of change of the IANA root is a limited resource, and the limit per year or recurring application period is about half of the number of entries of the IANA root at present, then the quantitative utility of any "EOI" effort is exhausted when a high confidence level exists for the number of applicants for units of change is less then, or more than, about half the number of the entries of the IANA root at present. Since every estimate offered by Staff since Paris has been several multiples of this figure, estimates which the "group of participants that engage in ICANN's processes to a greater extent than Internet users generally" have not dismissed as wildly in excess of the individual and collective estimates of those more engaged participants, it is self-evident that the "EOI" can serve no useful quantitative purpose. Thus, issue 1 "How do we ensure that participation in the EOI accurately represents the level of interest?" is misstated. The only useful question, assuming a limit on the resource to be allocated, is "How do we ensure that participation in the EOI indicates to the desired level of confidence that the number of applicants is less than, or more than, this limit?" Questions 2, 3, 4, appear to attempt to ensure a higher degree of accuracy than is necessary to answer the question of whether the limit is exceeded or not, and so impart no information, while proposing to alter the process to create an application closure prior to both ICANN's general outreach project and the opening of the application process. As no useful information can be gained by useless "accuracy", and profound anti-competitive and institutional confidence issues arise from ICANN itself gaming the rules to benefit "a group of participants that engage in ICANN's processes to a greater extent than Internet users generally", these questions are worse than useless. Since there is no useful quantitative information to be obtained, other than the theoretically possible result that there are no applicants, or fewer applicants than ICANN is technically capable of accommodating in any application period, we should examine what qualitative information is available. This is the sense of question 5. As the historic and ongoing "shell registrar" problem shows, there is little value in answers to the evaluation criteria questions 1 to 12. No clustering of origin or authorship, no non-uniqueness of the fundamental applicants, must be disclosed. A useful datum is how many, and how large, string contention sets will exist under the current proposed rules, and whether some of the proposed changes, not yet adopted by Staff, substantially alter the complexity of string contention and frequency of auction as the final allocator. While this can be done by asking applicants to identify their strings, this approach suffers from attempting to disclose closely held information, and from the reality that last minute applications from previously undisclosed highly capitalized parties are both possible under the existing rules, and more desirable than any restraint on competition that would bar last minute or patient applicants with sufficient means to prevail over insufficiently capitalized applicants attempting to claim some prior right, whether like IO Design, through a property claim, or like the recent ".shop" party at Seoul, through a marketing campaign. An alternative is to simply suggest some strings and request that applicants to indicate which string is like the one(s) for which they will apply. The result would provide a lower bound on the number of string contention sets, though in theory Staff's pick of strings could be exactly those strings known only to the EOI respondents, and upper bounds on the size of contention sets. This is the sense of evaluation criteria questions 13-18. It is difficult to identify how disclosure of any string set would critically assist ICANN after scaling the probable contention set in breadth and depth. Since the WG-C reported its "7-10" recommendation, ICANN has distinguished between applications which have only price as the policy, and those which have more policy than price. In DAGv3 the terms for these are "standard" and "community-based", though "open" and "sponsored" have been used for most of the ICANN period. One useful datum is whether the number of either of the "standard" or "community-based" applications alone meets the resource limit. This is the sense of evaluation criteria question 18, and 21. Evaluation criteria question 19 does not provide useful data not already provided. Evaluation criteria question 19 could provide useful data, if the rules made that data useful, which at present Staff has not found useful. Knowing the purpose of an application is of no utility if two or more members of a contention set based upon string similarity may not act upon that knowledge and submit alternate strings. This datum would modify the data derived from evaluation criteria questions 13-18. Evaluation criteria questions 22 (SLD reserved geo-names) and 23-44 (registry services description) and 45-50 (financials) provide no useful data not provided by evaluation criteria questions 18 and 21. As the RSTEP evaluations are conducted by an inelastic technical community, knowing just how many EOI respondents assert that they intend to trigger an RSTEP would be data useful to ICANN. That concludes the discussion of Question 5.As the objection processes are conducted by third parties, there is little useful information ICANN can obtain by learning how many EOI respondents report that they intend to apply for a string for which a third party right exists, or which raises public morality issues, whatever those may be. Thus, as interested as the potential objectors with standing may be in early identification of potential applications which harm their interests, that is not data which is useful to ICANN. Question 6, "Must the responder commit to go live within a certain time of delegation?" is slightly useful, in the sense that if a large number of applicants allocated a scarce resource defer their actual use of that scarcity -- one of a metered number of changes to the IANA root during some fixed period of time, a year or an ICANN new TLD application round -- then the "next" entry(s) may use the unused allocations. This does seem to be more theoretical than likely, and the point of the EOI is to establish useful knowledge with a useful confidence level. Question 7 is simply significant editorial review to ensure consistency of the explicit, and implicit process model, as well as revisions to, and public comments on, the new gTLD document set going back to Paris, and even Los Angeles or San Juan. Question 8 has been answered in paragraph 6 of this note.'... profound anti-competitive and institutional confidence issues arise from ICANN itself gaming the rules to benefit "a group of participants that engage in ICANN's processes to a greater extent than Internet users generally" ...' The outreach and marketing plan of record to be conducted during or after the last editorial round on the DAG is not sensible if application cut-off is planned for some date prior to the general availability announcement of application acceptance. Thank you for reading this small note.In a personal capacity, though I am employed by CORE, which has an interest in the issue. Eric Brunner-Williams |