And experience suggests that fields designed by individual
implementors aren't suitable for widespread use.
One obvious counterexample is the very widely used Delivered-To field,
which provides much more effective loop control than any IETF field.
A single counterexample doesn't refute the general principle, as long as
the generality isn't treated as if it were an absolute.
Feel free to submit an I-D that describes Delivered-To and request
publication of that document as an RFC. If it's really as good as you
say, it will get published and more implementors will be able to learn
about it and implement it correctly. Of course, if it's not as good as
you say, its deficiencies will be exposed.
Someone who makes the serious mistake of trusting your type of registry
might think that the name Delivered-To is available for a new field.
I doubt that anyone wants to reuse a field name which is already
in wide use, regardless of whether that use is beneficial or harmful.
It's certainly useful to list fields that are in wide use. I just
don't want the registry to become a means for either claiming new
field names or bypassing the standards process in establishing
legitimacy of new fields.
Keith