> One point that I need to make about RFCs is that, for a protocol to
> become standards-track, there is a process that needs to be followed.
> Part of this is that there need to be multiple interoperating clients &
> servers.
That's why I suggested (on the nnfs list, I believe) that we should
probably start with an "informational" RFC (once we have the design). Or
maybe "experimental". Those are relatively easy.
Once there is enough interest (and working code), it can then be put on
the standards track.
There's a RFC detailing all this, but I have no intention to go hunting it
down right now :-)
MfG Kai