Several Issues have been raised in the past couple of days which I
will try to answer here.
1) What exactly did I (The Chair) say in the last call.
Well, The WG last call for MNEMONIC was to publish both the Character
Mnemonics tables and the mnemonic charsets protocol together as a
"set". The draft "Character Mnemonics & Character Sets" was
recommended as an informational document and "Mnemonic Character Sets"
was recommended as a proposed standard. As a set, these documents can
reference each other and not cause problems. Many protocols,
including most routing protocols are defined in multiple documents,
which are editorially bundled to keep their references in sync.
2) What is a "recommendation" for an informational document.
Anybody anytime can publish an informational RFC provided the RFC
editor does not feel it is damaging or an end-run around the standards
process. An IETF working group can be chartered to produce an
Informational RFC, such as the mnemonic character set tables. In such
a case, the working group asserts before publication that it
represents a community effort, is believed to be accurate, and is
useful. This is a "stamp of appoval" for the ideas expressed in the
document.
3) Can an experimental document be referenced in MIME.
The working group has not made a clear statement on this question, but
it appears from a reading of MIME that so long as the character set is
documented in a publicly available form, preferably an RFC, it can be
registered. As such, and experimental RFC is an RFC for this
purpose. An Informational RFC would also be acceptable.
Now, it is worth mentioning that there has been quite a bit of
confusion about exactly what an "experimental" protocol is.
Fundamentally it documents the work of an ongoing experiment. It also
has been the destination for protocols not yet ready for proposed
standard. Because there are many protocols in use in a limited
community, and as such do not have the community push for
standardization, a new category of protocol is being proposed, called
"Prototype". This would be a non-standards track protocol which is
believed to be implementable and useful, although for a limited domain
of use.
Mnemonic #MAY# fall into the category for which this new protocol type
is being proposed. If the WG is unwilling to standardize the
protocol, and understands that this protocol is truely more than an
experiment, we could try to recommend it for a new kind of state.
4) Will the IAB accept a non-ietf character set or protocol for
reference in MIME.
While there has not been a clear statement on this question, I think
that the patern over the past few years indicates that stable,
publically avaliable standards are acceptable to be referenced and
used in Internet Standard Protocols provided the reference
unambigously points to a specific version of a specific standard.
The MNemonic documents #could# be published as Informational RFCs
republishing and documenting for the Internet community a Danish
National standard, or EUnet standard for instance. Once made
available in this fashon, it could be registered for use in MIME.
Note, this does NOT make the documents Internet Standards. An example
of this is Sun Microsystems republication of their NFS Standard. (RFC 1094)
Greg Vaudreuil
Chair
Internet Message Extensions Working Group.