Colleagues, I write this note provide one view on an issue discussed in Monday's NARALO monthly call. As a policy making body, ICANN shares a general characteristic of all policy making bodies engaged, for whatever reason, and where ever located, in making policy that pertains to the behavior of the vast collection of networks we call "The Internet". Its policy making is pre-scientific. Lots of policy is considered, some policy is made, but it is made without data, at least data sufficient in general to discern if the policy goal is implemented. The ability to correlate effect and cause is quite limited. Significant rhetorical resources are expended to maintain various claims of causation or necessity, but in the peer reviewed scientific literature on measurement and operations of networks, these claims are without foundation. I recommend reading kc claffy's two highly accessible works in this area: http://blog.caida.org/best_available_data/2008/04/16/top-ten-things-lawyers-... http://www.caida.org/publications/presentations/2009/top_ten_fcc/top_ten_fcc... ICANN's proposed registry agreement does not require new registry operators to provide access to operational data. ICANN's current registry agreements, all 17 of them, also lack provisions to require the operator to provide access to operational data. That's about 3/4ths of the DNS, higher if the root constellation operated by CNNIC and its data is excluded, for which the only data ICANN desires is contained in the registry operator monthly reports. Additionally, ICANN has never requested recursive resolver operators such as (US): SBC (AT&T), Comcast, RoadRunner, Verizon, AOL, EarthLink, Charter, ... and (CA) Bell, Shaw, Rogers, Telus, Videotron, ... to provide data to researchers. Further, ICANN has never requested browswer (stub resolver) vendors to provide data to researchers. A vast amount of data gathering is being conducted, but for user profiling purposes, to sell shoes, but not to ensure that we know the shape of things to come, and can distinguish policy successes from policy failures. Why is data necessary? In October 1986 the NSFnet phase-I backbone dropped three orders of magnitude from its capacity of 32 kbit/s to 40 bit/s. The cause was not predicted. More is possible. We have no means to associate probabilities to continued operation, or failure. The proposal I made was simple and clear (I thought). Most of the data is in North America, and subject to the particularly legal culture of the United States and Canada. To improve research access to data, and therefore the ability to conduct network policy informed by data, rather than pre-existing beliefs, which may as well be religious articles of faith, ICANN's StratPlan for 2011-2014 could be changed to make reducing the legal barriers to data a goal. The ALAC comment on the StratPlan started with the text I suggested, to the effect that most of the crucial data is "in North America", but concluded that the plan should have increased regional diversity as a goal. Rephrased as an either/or, either a goal of a global decrease of legal restraint on critical data, or a goal of regional replication of legal restraints on critical data, are the choices. I proposed the former, and as was mentioned in Monday's NARALO call, ALAC chose to substitute the latter. I hope this makes clear one view of the issue discussed on Mondays' call. I appreciate that there are other views. Eric