Re: PGP fingerprints in CREDITS file?

Matti E Aarnio (mea@mea.cc.utu.fi)
Wed, 15 May 1996 22:16:01 +0300 (EET DST)


Linus in venting some heat (we had a brief spell of summer with +25C hot
weather here arounds -- that's a lot to us frozen finns.. Another reason
for having a Penguin as a mascot :-) ) about MIME:

> Not only don't I use PGP, I don't want PGP-signed email, and I don't even
> want MIME-encoded email.
>
> Too bad Zimmerman wasn't convicted - he wasn't guilty of breaking any
> silly cryptography laws, but he _was_ guilty of very bad programming. The
> PGP single-character escape stuff is broken, in bad taste, and should not
> be allowed (*).

I agree, and PGP-in-MIME has not emerged to take over that
properly. (PGP was made about at the same time as MIME, and
without any contact between the two groups at how to do the
quoting + encoding in ASCII transport..)

> The people who created MIME not only should be convicted, they should be
> shot on the spot.

Tut tut! I am glad that here is no NRA (National Riffle Assoc)
lobbying for the profileration of guns -- these flamboyant opinions
might turn into real thing othervise.. Here, have some Herring ;-)

Reasons for the design decissions were multiple, most important
was "the message must be transportable over braindead systems
existing in the network", which made rather strict 7-bit ASCII
requirement, and REJECTED UUENCODE because it can't be sent
without propable corruption at ASCII/EBCDIC transition borders..

We "MIME-people" see it as "not ideal, but pretty good for the
junk it has to live with". See X.400, and you may start to
realize, why instant "scrap all old, replace with new" approach
was at very early decided unviable.

Of course there are ways to optimize, when MTAs have online,
or offline agreenment on "extended capabilities" (like 8-bit
transport), they can avoid conversion into "braindeath resistant"
format.

> I have a reasonably mime-aware mailer (pine), but I
> _still_ have more problems with MIME mail than I ever had with non-MIME.
> I'm seriously considering refusing to even look at mails or patches that
> use MIME-encoding other than cleartext.

I am slowly giving up at getting people to use proper terminology,
the "mailer" is system transporting messages between systems, it
is not USER PROGRAM. While I pretty much abhore X.400, they at
least have the terminology well in hand (but they got it from
IFIP in 1970ies to begin with..):

MTA = "MAILER" = Mail Transport Agent
MUA = "pine"/"elm" = Mail User Agent; "Mail client"
MS = "/var/mail/torvalds" = Message Store; "Spool"

So, you have reasonable MIME-aware mail-client, but not perfect ?
Which version of PINE it is ? The lattest ?

I have mime-semi-aware mail-client, but my MTA (the mailer) knows
what to do with trivialities like QUOTED-PRINTABLE in message,
when storing incoming messages into my message store ( = mailbox )
(I do not get any QP stored into my mailbox, it will always be
decoded to 8BIT.)

> And don't tell me "MIME works fine" - I get quite enough emails, thank
> you very much. It doesn't work fine at all.

It depends, usually when I have trouble with MIME email, the
sender has misconfigured his/her client, and produces flawed
messages. ( = non-MIME email disguised as MIME.. )

> As it stands now, I have to edit the damn thing by hand or install a program
> that I don't want, and this results in me mostly ignoring the damn broken
> thing. Yes, I'm grumpy, and that's because I just I got _two_ mime-encoded
> mails that didn't even uudecode cleanly before I edited them by hand.

Tough.. I too might feel grumpy after such.
(But I do have "mpack"-toolbox at hand too, it can do wonders..)

> Linus

/Matti Aarnio <mea@nic.funet.fi> <mea@utu.fi>