On Mar 20, 2013, at 3:25 PM, David Farmer <farmer@xxxxxxx> wrote: > "xxx is obligated to ..." wasn't intended as a suggestions for text, but like I paraphrased the text from the draft above, and I intended it to paraphrase the the text that needs to be added. The text above quoted from the draft "...such recommendations must be taken into consideration..." and I agree with that. I'll note the care in how it is worded, but it seems to flow only from IETF to IR System. The IETF defines the protocols, including the parameter fields. While the application of the protocols is up to the community, the IETF does also on occasional state known technical recommendations or constraints in usage of its technical standards. The original RFC2050 specified such technical constraints, the RIR/operator community has generally respected them over time, and hence they've been included in the revised text. > Maybe I'm just being overly sensitive to the "must" in "...such recommendations must be taken into consideration...". But, it seems to me that this says IR System must defer judgements on technical issues to the IETFs technical recommendations, even if they are old, crufty, and been left in dust by the march of technology. We have existence proof that the Internet Number Registry system is quite capable of determining when circumstances have changed and while generally considering important advice, e.g. hierarchical assignment, the registries have indeed evolved policy as necessary and specified by the community. Note also that that the actual text does not at all say "defer judgements on technical issues to the IETF's technical recommendations", but instead that those recommendations must be considered in policy discussions: "In addition, in the cases where the IETF sets technical recommendations for protocols, practices, or services which are directly related to IP address space or AS numbers, such recommendations must be taken into consideration in Internet Numbers Registry System policy discussions regardless of venue." I understand that you're seeking assurances that the IETF will dynamically update any documents that affects the Internet Number Registry System, but reality is that it takes specific efforts to make that happened (and they don't always converge with productive output.) This effort to update RFC2050 is at least a step towards getting better alignment than we have today. FYI, /John Disclaimers: My views alone. No header fields were harmed in the making of this email.