Eric,
I'd suggest that this distinction be made much more explicit: list the
raw mode in a "standards' track" RFC, move the tagged mode to an
"informational" RFC (or remove it entirely which is my, and several
other people's, preference) - if you don't do this, you will be creating
a whole lot of marketplace confusion and will neither be "facilitating
interoperability" over anything more than a very short timeframe nor
making "good standards" (as in "good standards need to describe how to
live with existing implementations of things, some of which are the
product of different traditions" - Harald Alvestrand). I am assuming
that Harald did not mean to endorse the practice of implementing, and
shipping in volume, implementations of early-stage IETF drafts, and then
using any momentum built up through those means to force inclusion of an
option in the final IETF standard (though that would be a "different
tradition", of course, from the good old IETF way of doing things).
Andrew Smith
humm , wouldn't IP fall into this category ? V6 was built upon V4
implementation experience.