Karl, I accept that in your view Depository is attempting to innovate and experiment. I don't think it is. Absent the suggestion by McTim that this Advisory Group should form a view on the ASO process, or the appeal to an untested authority inherent in ICANN through "hierarchy" offered by Mike Burns, the substance and process are appropriately addressed on the ARIN-PPML list, in which you are free to participate. Turning to the history of RIRs, two mechanisms intended to scale as the price of network adapters and of network attachment dropped have been adopted. I proposed a regional scheme, based upon the ITU's X.121 Recommendation, specifically the zone, represented as the first octet of a Data Network Identification Code (DNIC) to create a small number of regional registries. The ISO 3166 scheme was adopted instead, creating two orders of magnitude more registries, for name to address mappings. The model of a regional scheme was adopted for address allocation registries, as much in recognition that no specific competency existed in North America which did not also exist in Western Europe and Japan, as for territorial diversity, and the attendant jurisdictional diversity. If consolidation were seriously proposed, it would be as arbitrary now as it was earlier to assert which of {ARIN, RIPE, and APNIC} is the first RIR to be eliminated. I don't think the characterization "For no technical reason, but for every political reason ... created two more RIRs" is correct. There is technically defensible reason for delegation in technical institutional development, a course pursued with diligence by the Network Startup Resource Center, and others. There is technically defensible reason for non-revocable delegation of address resources, as Geoff Huston's projections show, the AFNIC and LACNIC regions will be capable of general allocating v4 resources for several years after the APNIC, ARIN and RIPE regions transition to exhaustion allocation policies. The aggregation is an enemy of regional registries claim would be more credible if, as a general rule, access networks were predominantly trans-regional. As they are not, for pricing and tariff and other jurisdictionally scoped policy reasons, this is a weak claim of non-necessity or inefficiency. Finally, assuming all of the initial claims of Depository for the purposes of argument, there is the specific choice of replicating the domain market structure, in the address market. There are two final claims to consider: first, that the domain market structure is good, and the second, that innovation and experiment are actually less useful than replication of an existing structure. I don't think either of those are reasonable claims, or capable of proof, though I understand they are offered as proved by assertion by Depository, in its own pecuniary interest. Eric