Oh, well, looks as my credibility is at stake...
I have consulted technical sources within EUnet about precisely who
has installed Keld's software and what it does. The information was
enlightening to say the least.
His software is being installed because it provides an 8-bit clean
SMTP implementation. That is their only interest in it -- as a
mechanism to transport ISO-8859-X cleanly between their backbone
sites. (EUnet has mostly a hub and spoke topology within each country
and one backbone hub site per country.) The EUnet folks I talked to
were all very firmly of the mind that 8-bit SMTP was essential and that
their users all felt it was essential.
I do not beleive that backbone managers have installed it only
for the 8-bit SMTP. Well, anyway it would be interesting to see
how and why they have installed it, I did not help each and every
of them to do it.
But the 8-bit part is of cause important when you are addressing
customers that run 8-bit clean, and then to turn their mails
into RFC-compliant mail. This is what I see as "consenting
adults" agreements, and gatewaying from the sexy 8-bit
sites out in the real world. We have something like 20 sites
in Denmark running their connection to us in 8-bit clean mode.
I have never heard of an EUnet policy of generally doing 8-bit
SMTP, which is what you are indicating? Or even between the backbone
sites. I think you have misunderstood something, Ran. (Or maybe
you have been misinformed).
There apparently is agreement by most of the hub site administrators
to implement Keld's software ("patches to sendmail/SMTP" to quote one
note) soon but many have not yet done so. The patches have reportedly
been made available to Paul Pomes of UIUC for inclusion in the version
of sendmail that he maintains.
Well, mnemonics is adopted as EUnet policy, as you also saw
our chair, Michael Nowlan state yesterday. There are minutes of
the meetings where these decisions are taken, and I am now allowed to
quote them. It will take me a little while to find them, but
I expect to have these minutes exerpts ready later today.
I have stated earlier (something like half a year ago) that
it is not all of EUnet backbones who have installed the patches.
There were no statements in support of using Mnemonic for anything.
All of the comments were about Keld's wonderful 8-bit clean SMTP mods.
Well, that is the same thing. You do not run just 8-bit, but we
run in a mixed 7/8 bit world. But again, I cannot tell how people
have perceived the software. I have never presented my patches
as just an 8-bit clean way of doing SMTP, but clearly stressed the
mnemonics part of it. maybe they do not know the word Mnemonic,
but the codeword in EUnet is keld-char. I did not invent that, but
obviously I am the inventor of it - so that is how they dubbed it.
On another note, I heard from someone in the Nordunet Engineering
Task Force (NETF) who indicated that Mnemonic was originally put in a
draft of their mail document mainly because Keld had claimed that
other organisations were in favor of using it and had committed to it.
Yes, that is true, and that is also a fact. I was not the only
person from EUnet at these meetings, and the others could have
corrected me if I was misrepresenting facts. This also include
RARE people.
When a NETF member investigated, that person was unable to
substantiate the claims made and I'm told that Mnemonic has now been
removed from the NETF document. I have not heard this from anyone who
is an NETF officer but I do consider the source reliable for my own
part.
Oh, well, such stories... I do not know which documents you are
talking about, but the NETF meetings on the issue I have heard
of, which accounts to 4, and with 2 meetings where I was not present
at the time of the decision, all concluded that they preferred
mnemonics.
I infer that the IAB and IESG have not heard formally from either
EUnet or Nordunet in support of Mnemonic since none of the Chairs of
these working groups have reported same to any of the lists. There
were formal requests for support of ISO-8859-1 and 8-bit SMTP that
reportedly were received some time back.
Well, there is at least one NETF MHS minutes which states that
mnemonic was the preferred interchange format, sent to the
822ext mail list.
And there is the EUnet chair statement of yesterday saying
that mnemonic is EUnet policy.
Of a less formal nature, you have the chairs of eunet-char-wg
(namely me!) and NETF MHS (namely Harald Tveit Alvestrand)
stating requests of mnemonic. I do not think any of these
statements were stated as on behalf of their respective
organisations, though.
My fax number is +1 (202) 404-7942; please mark faxes to my
attention as it is a shared fax machine. My snail mail address is
below. If anyone has an official document from either EUnet or
NorduNet that is recent and on official stationary that documents
these official endorsements of Mnemonic, feel free to fax them to me.
Rest assured that I will telephone and fax back to the persons
responsible in those organisations for confirmation of the current
facts before reporting on any such faxed documents to any of these
lists. If I don't receive anything, I will assume that the claims
that support for Mnemonic is officially requested of the IETF aren't
correct at the present time.
I agree that there is no formal written statements from EUnet nor
Nordunet to IETF on the issue. I can ask the parties if they
want to do such a statement. But nowadays we do our administrative
things by email... Do we need to be that formal with signatures and all?
I think the formal address should be that of the 822ext chair,
Greg Vaudreuil, tho.
Ran
atkinson(_at_)itd(_dot_)nrl(_dot_)navy(_dot_)mil
Randall Atkinson
Mail Code 5544
Naval Research Laboratory
Washington, DC 20375-5000 USA
Keld