Sorry, just one more message from myself, because Murray Kucherawy
said in a moderator message exchange

  Nobody is asking that you leave.

After i said it was my last message (except that the draft needs
one more iteration).

But in the meantime i have read (via web archive) that it was said
that the other delta algorithm is a good thing.  And noone spoke
up against that.
So then i do, one last time.

The other approach is not good, despite all the non-real-life
and blindingly dressed up examples that always can be seen.
My first remark on message (part) reencoding from when it was
posted first is still valid, and will remain so.  And this is not
only about entire MIME parts, it is also for individual headers.
For example, the IETF lists themselve post, without any technical
reason,

  Subject: [art] Re: [Json] ECMA and JSON Schema
as
  Subject: =?utf-8?q?=5Bart=5D_Re=3A_=5BJson=5D_ECMA_and_JSON_Schema?=

which is maximally suboptimal in several aspects (encoding not
needed, plain 7-bit ASCII encoded as 7-bit clean 8-bit UTF-8).
Now, if a technically not totally worn down MUA responds it'll say

  Subject: Re: [art] [Json] ECMA and JSON Schema

without any MIME at all.  Given that so many Python based software
is on the way, which "D(ocument)O(bject)M(odel)" any emailish via
the Python email machinery, i would bet such changes happen
naturally.  It is a bet, but a good one.  I won't test now.

But in reality changes happen much more often and for much more
trivial conditions, compared to the blindingly good examples.

In addition content needs to be checked so that the delta
algorithm syntax does not clash with header content, so that
possible base64 (i think) encoding can be applied (i think).
Each and every snippet.  And there can be plenty.  Then possibly
must be encoded for the reasoning of syntax protection only.
(And if only comma, equal-sign, semicolon, i have not looked at
latest variant.)

Please compare this with what the mailman[23] mailing-list manager
does, and with all the headers that players like Microsoft and
Google produce, and things like IronPort.
These headers (and parts) are practically all base64 per se.
I am surely not that wrong when i interpret this in parts with
"why spent a single thought, just go the safe way, period".

Again the IETF comes over with an overengineered and overly
complicated approach to a simple problem, with a new algorithm
noone has asked for.

Compare this fact alone with the ACDC approach, which makes use
of Colin Percival's BSDiff algorithm, which is excellent and
its core being used by so many parties and purposes for such a
long time.  And simply stores this compressed, and always encoded
as base64.  It may be a little bit larger due to the always
required base64 overhead in not few cases, but as soon as real
changes have to be tracked, things change quickly.

And then this is always the same, sort the headers in stack order,
as you need to for generating or verifying a DKIM signature, and
pass it into the algorithm, and you get back the same, in and out.

Maybe later, in all all-8-bit-Unicode email environment, such
things do no longer matter.  But this likely will take decades.
Despite the many languages which pay a size-overhead price for the
always necessary encoding of their character data.  S/MIME and PGP
require encoding due to normalization patterns anyhow.  MBOX file
content protection based "From_" quoting can only be circumvented
by applying a MIME encoding, too.  I do not know.

In my opinion thus the delta algorithm of what the IETF wants to
adopt is worth of criticism.  *Too*, that is, *too*.

So, but now, except for another iteration of an error
i introduced, as not expecting any discussion anyway, and hoping
that the next IETF meeting plan will not include my name no more,

Good bye from Germany.


P.S.: we need SMTP/TLS, except for TUHS *all* connections i have
is via TLS, discovery via SRV is an easy thing to do, and native
to SMTP servers, as they need MX lookups anyway, etc etc etc.

--steffen
|
|Der Kragenbaer,                The moon bear,
|der holt sich munter           he cheerfully and one by one
|einen nach dem anderen runter  wa.ks himself off
|(By Robert Gernhardt)

_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to