Charles,
your message to the IETF-822 list needs some comments - and updates -
from my side. So I'm sending this reply to you (and perhaps it appears
on the list).
On Sun, 23 Feb 2003 16:36:07 GMT, you wrote:
Yes, I have been round this before. It is clear that, even if the Yenc
people were to use
Content-Transfer-Encoding: x-yenc
which they could start doing tomorrow, things would be much simpler.
For whom would be things "much simpler" ?
I have tried to persuade the relevant people to produce an internet-draft
proposing it as an official CTE. But they don't want to know.
Please realise that "they" (me) always wanted yEnc embedded into MIME
and that "they" (we) collected far more information about the entity
of problems than anybody else.
btw.: The work is "in progress", not burried.
The latest proposal from Russ Albery (posted to news.software.nntp)
is a part of that work.
My latest problem is that the "Content-ID" which is used to identify
multiparts can be easily _faked_ on Usenet. News people have far more
serious problems with MIME than the Mail people can imagine. The
solution to that problem is far more complicated and complex than yEnc
itself :-(
They seem
convinced that the IETF is totally against all change, and that they would
be heavily sat upon, and that the whole process would be far too
cumbersome and time consuming,
If I'm looking at the latest discussions about USEFOR here then I
really dont want to imagine what could happen to a yEnc RFC.
And there is one thing you should never forget:
yEnc is a proposal for transporting binaries _efficiently_ and
crc-protected of USENET. As soon as it is a MIME encoding this entire
stuff would spread to eMail. And I would be really curious how many
mail-relais are corrupting 8-bit-posts then. Furthermore existing
eMail clients seem to _modify_ the body - especially when it comes to
leading/trailing tabs/spaces (but also ascii-dec-26)....
Usenet is a perfect playground to test and evaluate such toys as yEnc.
But eMail is a _business_critical_ application around the world.
I believe that this is the reason why this list (and IETF) is
extremely careful when it comes to changes for eMail.
So please apologize that I dont want to be one who is messing up the
worldwide eMail system 'just to save 25% of bandwidth' - which is
negligable for Mail - but important for Binary Newsgroups (where yEnc
belongs to).
and "hey it works fine already, we don't need no stinkin' IETF".
Yes, yEnc is working pretty good. It is saving Gigs of bandwidth every
day for USENET.
FYI: Usenet does not need IETF. The neticens there have invented many
nice things on Usenet even without the admins and us - the developers.
These people dont _care_ about us (including myself).
All very sad, but that's how some people regard
the IETF. So we are stuck with indicating the presence of yEnc by a flag
in the Subject header. It is a mess.
Update: More and more posts appear on Usenet even without the
mandatory keyword "yEnc" in the subject line. It is necessary to scan
the body.
More and more binary posts appear on Usenet with invalid filenames. It
is necessary to scan third party systems and protocols to know what's
inside a binary.
More and more binary posts appear on Usenet without a subject - or
with a misleading or "encrypted" subject (uncle tom arrives tomorrow
at 7pm on JFK International)....
Usenet is not eMail.
And Binary Usenet is not Usenet.
The _real_ good question would be if it makes any sense to TRY to have
them compatible. From my latest experiences I'm even convinced to say:
Fingers Off.
Last not least:
Introducing MIME in the Binary Usenet is one thing nobody ever was
able to do. All that "nice stuff" which is possible with MIME is not
what neticens expect there - or what they want. Together with the fact
that all newsreaders (and all news->web gateways) must be modified
then it seems to be more than hard work. And because the neticens
expect smooth transition (the posts must be at least usable on 99% of
the newsreaders by the help of external decoders) the yEnc embedding
into MIME must be realised carefully.
This requires also a proper check of _all_ the MIME capable
news-readers out there. And I'd really like to know who wants to do
that job..... (It takes years of hard - unpayed - work).
That's my current status - 15 months after the publication of yEnc, 24
months after the first discussions about it.
Today things are more difficult than yesterday.
CU
--
Juergen
--------------------
www.yenc.org