On Mon, 2007-11-05 at 15:13 -0800, Darren Reed wrote: > Or do we simply say that the project defines whatever it wants > to be its spec, use the internet draft as a reference and if/when > the RFC becomes published and there's a change, someone > hopefully remembers and publishes another case that updates > this one?
This is our existing practice. (IPsec NAT traversal is a good example; the wire protocol didn't actually freeze until the RFC issued).
