Donning my time machine paraphernalia and asbestos suit ...
Is it possible to analyze these problems by asking first whether ICANN has imposed these requirements (for e.g. a stakeholder council) on the applicant as a condition of entry into the root or, instead, it's a voluntary commitment undertaken by the applicant for its own purposes (e.g. to get the support of global environmental organizations, to get external funding, or what have you).
This is asking for a trip back into Kurt's mind, from San Juan to mid-'10, as we went back and forth on the then-new alternative to the form we'd had in the 2000 and 2004 rounds where the applicant offered some form of extra-contractual policy oversight as "sponsorship". The several versions of the Applicant Guidebook captured several different instantiations of the proposed criteria and how applications meeting the criteria-de-jour benefited -- generally in anticipation of contention with contemporanious generic applications, and suffered -- generally in restrictions unlikely to motivate registrars to incure the (hypothetically) recoverable cost of coding and staffing the conditions placed upon would-be registrants. So, did "ICANN impose requirements" such as stakeholder councils? For the 2010 cohort of applications which self-identified (and read the rules) as "community based" application types the answer is "yes", though of course the necessary fiction was that we agreed to all this, <insert the usual "bottom-up, consensus" boilerplate here>. For the 2000 and 2004 cohorts of "sponsored" applications it was the applicants which came up with their own form of sponsorship. We don't of course, know how Staff is going to evaluate the "community type" experiment. I've heard (back in '09) that no one then remembered the 04 or 00 set of "sponsor" issues, and of course it was obvious that (pace past employer) none of the 2000 cohort, sponsored or generic, were stand-alone successes, and of the 2004 cohort, only .cat could claim to be a stand-alone success, and consequently, some of the 2004 "sponsored" type cohort were trying to survive as generics. The next "Kurt" (and supporting staff) could change the deck chairs again. Going to the condition-of-entry-vs-voluntary-commitment question, for the 2010 cohort, some applicants anticipated their string being placed in a contention set with the string(s) of other applicants account and filed "community type" applications. Some applicants, those with cultural and/or linguistic and/or municipal and/or regional service models, filed "community type" applications. So you have those that gamed the contention process, some successfully, and those that would have reasonably obvious sponsorships had the 2010 rules been, on this point, similar to the 2004 and 2000 rules.
If it were the former, then I think I do have a problem with ICANN's enforcement of the provision absent some demonstration both that it is reasonably necessary for the stability/security/interoperability of the DNS (which I wouldn't think it is, on the face of it from your hypothetical) and that it is otherwise within the confines of the mission.
You have to reconcile your preference for non-enforcement post-delegation with the benefit acrued during contention set processing. Similar reasoning appears to have motivated some applicants to write PIC statements, as there was pre-delegation benefit. <insert a generic wish for robust registry and registrar compliance here.>
On the other hand, to the extent that the applicant made some voluntary undertaking that was not viewed by ICANN as a condition of its entry into the root, I have less of a problem in saying that ICANN can take steps to enforce the contractual promises made to it by third parties.
Modulo the absence of post-delegation compliance (see .pro, .travel, ...) you're now addressing the set of applications which had no expectation of string contention, and sought no benefit from a PIC or a "community based" type of application, e.g., .scot, .berlin, .lat, ... So this particular "other hand" is empty, or close to it.
I know it's a fuzzy line - maybe too fuzzy to be workable. But I do think that it might be the line that we've been struggling to define throughout this discussion. David
IMHO, it means opening the gambler's intent can of worms, and co-mingling those attempting to game the system to gain delegation, and those who just made ability to read simple instructions in Catalan (for several values of Catalan) the predicate condition for registrants seeking to register a name in a particular namespace. Eric Brunner-Williams Eugene, Oregon