I appreciate your reasoned and sensible reply to my impassioned
and frustrated note.
It's precisely because the PWG members *hope* that the IESG
will allow/recommend the advancement of this new Printe MIB
revision to 'Draft Standard' status that I and my fellow
implementors and designers at the Xerox companies wished to
nip the cancer of 'ASCII-only' for any string object not
covered by 'prtGeneral[Current|Console]Localization' out of
this draft and gain the latitude for other non-English,
non-ASCII marketplaces to use a 'reasonable' character set
and language/culture locale for their implementations.
Historically, once a MIB goes to 'Draft Standard' state,
nothing substantive is ever done to it again (because
the IETF conceives of 'Draft Standard' state as early
warning that the exact text may SOON advance to '[Final]
Standard' with some level of implementation recommendation).
I respectfully suggest that the text advanced to the IETF
next week (with any typos or formatting problems corrected
on the advice of the RFC Editor) already IS the Printer
MIB II - don't expect that there is another chance to
repair deficiencies if 'Draft Standard' state is granted.
The localization clarifications which I suggested were
essentially verbatim from Xerox/FujiXerox/RankXerox
joint corporate MIB corporate MIB standards (Tom
Hastings and I did the ASN.1, but a group of over
100 engineers worldwide contributed their comments
on the localization control objects). If we hadn't
had EXPERIENCE with this already, we wouldn't have
bothered to advance the suggestion.
A 'more global solution' than just the Printer MIB
is very unlikely - the SNMP MIB side of IETF standards
are at the tail end of recognition of localization
concerns.
Please note, that our XCMI (Xerox Common Mgmt Interface)
working group does NOT suggest that many (or any)
managed devices should bear the burden of message
and code translations - that properly belongs with
the individual client (management station) software
- all we wanted was no ambiguity about the full
POSIX.2 locale of the other string objects in the
Printer MIB (exactly as already present for
the 'xxxDescription' objects in RFC 1759) so that
the client software could do its job - which is
precisely the main recommendation of RFC 2130
(report of the IAB Character Set Workshop) for
all new APPLICATION protocols. Note that in MIBS,
SNMP is just a common transport. There is NO
chance that some future version of SNMP will
directly address localization in the protocol
(any more than that some future version of TCP
will do so). And since SNMPv1 remains the
IETF's official standard and SNMPv2 (in its
many non-interoperable dialects and versions)
is only a 'Draft Standard' (without any
administrative framework of substance), don't
bet on SNMPv3 dealing with locale specification
in the forseeable future (and supplanting the
ubiquitous SNMPv1 'least common denominator').
By suggesting that the good example of RFC 1759
be generalized into the HR MIB (RFC 1514), I of
course meant that the HR working group experts
(who may indeed reconvene) should take it as a
good example to start from. An HR MIB solution
would offer the right whole system and/or
subsystem scope for such a solution.
I'm afraid I personally have very little to offer
in the way of useful skills to the PWG in the
future. Before coming to Xerox 30 months ago,
I had never known ANYTHING about printers and
(largely) I still don't. But I do have 25 years
of communications protocol design and development
experience and about 15 years of open systems
management experience (more with OSI CMIP stuff
than IETF SNMP stuff).
Except as a conduit for the input from real domain
experts from the Xerox companies on a Finisher MIB,
I really don't have any basis for opinions - I still
don't really understand how xerographic printing
works.
In all likelihood I will be gone from Xerox by
sometime later this calendar year (if experience
is any guide, because some massive reorganization
happens and I'm summarily fired - that's the way
all of us independents who've moved past code-
slinging usually go). At the time, I will also
cease to observe or contribute to the PWG.
Good luck to you all in your efforts to advance this
revision of the Printer MIB to 'Draft Standard'
state - not very many 'Proposed Standards' every
get there - and almost NONE ever get to '[Final]
Standard' state.
- Ira McDonald (outside consultant at Xerox)
High North Inc
PO Box 221
Grand Marais, MI 49839
906-494-2434