In thinking about the new registry services PDP that is now beginning in the GNSO, I'm trying to find use cases for the planned decision-making process, and to figure out why the process should produce what result -- and how the rationale for the suggested decision can be cast into an objective criterium. I'm less interested, at this point, in arguments about the contractual standing that ICANN might have to deny or acknowledge any particular service.
Two examples tonight: SiteFinder and WLS.
I feel that SiteFinder is relatively easy -- not so much because of stability considerations, but most importantly for two other reasons: (1) It moved decisions from the edge to the center, thereby negatively affecting competition. (2) It caused cost to the public, for the benefit of a single player, much like environmental pollution.
WLS is a more difficult case. It grew out of a "tragedy of the commons" situation, as Rick Wesson observed accurately: Registrars were using up shared resources (transaction bandwidth) that came without a price tag to the point where these resources degraded, and "normal" registration was affected; using overflow pools and other limitations only moved the cattle to a different pastry. The WLS proposal appeared to introduce an approach to the registration of expiring names that would not lead registrars to engage in behavior that is detrimental to the public (although perfectly rational for the individual registrar!). The downside is that WLS gives precedence to a specific business model for registering expiring names. Different business models are essentially pushed out of the market.
Had there been an alternative to WLS in order to stop the detrimental behavior? I believe yes -- namely, changing the price structure for using registry resources in a way which rewards successful registrations, but financially punishes failed transactions and excessive use of bandwidth. In such an environment, publicly detrimental consumption of transaction bandwidth would be irrational (and ruinous); different providers of expired domain name registration could continue to compete on an equal footing, and with the business models they desire; many of the complexities of WLS would simply be unnecessary.
Now, how should ICANN have dealt with the situation in the light of this argument (which, I believe, wasn't made at the time -- but I might be wrong)? My current inclination would be to say that ICANN should have denied WLS, and should have favored the decentral approach. If that's indeed the right approach, then how can the argument be cast into a more generic form that would be useful for an evaluation process?