--On Wednesday, June 19, 2013 19:43 +0000 John Levine <johnl@xxxxxxxxx> wrote: >... > As a concrete example, the EPP systems used in production by > TLD registries use extensions that are documented only in > I-Ds, often expired I-Ds, or in dusty I-D like web documents. > If you look at the applications for new TLDs on the ICANN web >.. > Assuming we care about stability and interoperability, > wouldn't it make sense for the IETF to spin up a WG, collect > these drafts, clean up the language, make sure they agree with > the widely implemented reality, and publish them? I hate to put a damper on this sudden enthusiasm, but we presumably also care about issues of change control and adoption. It seems to me that we could reasonably do two things: (1) Try to do what you suggest but with the understanding that those who haven't come to us and said "standardize this" may still not come... and may not consider that they have any interest or obligation to conform, long-term, to whatever we come up with. I think that gets us Informational documents that are basically a cleaned-up version of the I-Ds. (2) At least partially follow the lead of Joe Abley's comment. Go to ICANN, suggest that having these commitments in the new gTLD process depend on I-Ds that are not supposed to be referenced normatively and might change at any time is not in the best interest of anyone, especially the best interest of Internet stabiiity, usability, and interoperability. See if we can persuade them to initiate a policy development process that would require adherence to whatever consensus standards are developed in this area. Then, per Joe's comment, let their staff poll the various actors to find out what they are doing, be sure that the various implementations are at least intended to be consistent and get that documented. Then we work with them to get the I-Ds updated and to work a process that includes reviews from those who are actually using and depending on the specs. Seems to me that would both minimize IETF workload and get us a higher quality and better-vetted set of specs for which there was actually a committed audience. john