Mathieu, Thank you for the suggestion, however ... The authors and the various eventual applicants to ICANN in 2000 and 2004 and 2010 for delegations for linguistic and/or cultural and/or municipal and/or regional service registries, as I recall them did not, in general, frame the agreements they proposed within their eventual applications as being made for the purpose of competitive advantage in a contention process. The authors of the applications in 2010 for delegations other than those for linguistic and/or cultural and/or municipal and/or regional service registries, again, as I recall them, rationally evaluated whether their business models could be achieved under the (presumed enforced) restrictions of the "community-type" (which bobbled through each of the revisions of the draft AGB, in form, and in dispositive function). My point being that there are "voluntary commitments" such as services to users who prefer a Catalan namespace, or a cooperative-operated registry, absent any application impediment such as string similarity. Dissimilarly, there are "voluntary commitments" which appear to have no meaning outside of the application type restriction of 2004, and more recently, the similar string contention resolution mechanism of 2010. We can distinguish between these, and should, as the exercise of "voluntary commitments" for the purposes of gaming type restrictions (2004) and/or gaming contention (2010), substitutes (gamed) implementation for the policy of record.
And most importantly, regarding the discussion about "voluntary commitments", as Avri points out, we might have a way forward if we were to agree that *the scope of acceptable commitments in any agreement should be defined by policy* (with all the related process safeguards, including bottom up nature as well as advisory inputs), instead of implementation. Then it would be up to the policy makers to define whether eligibility conditions are appropriate or not and should be enforced, whether a specific form of stakeholder consultation or governance is acceptable, etc.
As a process coda, prior to the June 2010 meeting in Bruxelles, there was no "single user" application type in any draft of the AGB. Our current problem contains the implementation-trumps-policy warts of the 2010 round, and we cannot pretend that policy has been made transparently, accountably, and from the bottom-up, though we can hope that it is in the present, and will be in the future. Eric Brunner-Williams Eugene, Oregon