From: Brad Knowles (blk@skynet.be)
Date: Wed Jun 02 1999 - 13:25:14 CDT
At 10:43 AM -0700 6/2/99, Brad Templeton wrote:
>> When it comes to processing signed articles in this fashion,
>> obviously everyone has to do the same thing.
>
> But they won't. We will have a transition, as sites slowly, over the course
> of years, adopt handling for signed articles.
Sure there will be a transition, but assuming we do our job
properly, there will be one and only one process for signing all
articles and dropping those that aren't properly signed.
Therefore, assuming that we do our job and write a good enough
standard and all the implementors do their job and they are actually
compatible with each other, there shouldn't be a problem.
> There are any number of headers that can be faked to cause trouble if
> you don't require they be the same.
Fine, so you sign all those, which includes the implicit
instructions to remove all non-local/non-variant headers that aren't
replicated within the message body.
In a signed document, the lack of a particular header is every
bit as significant as the presence of a particular header.
Unless you're arguing that there are variant or local headers
that could be duplicated and cause problems, but if that's the case,
then either signing standard would be subject to the same problems
because you explicitly do not sign variant or local headers.
> It also has no mechanism for a double signature, except the fairly ugly
> multipart within a multipart.
If that's what it takes to get wider adoption and implementation
of the same MIME standards, then I'd be willing to pay that price.
We're not talking about a huge volume of messages here, only on
the order of 500,000 to 800,000 today, which we know for a fact can
be relatively easily handled on the mail side by properly configured
single individual machines.
In terms of message handling load, I don't see how this would be
significantly different for Usenet news. This is a problem that we
can throw more hardware at.
However, to create a whole new standard from whole cloth to
accomplish largely the same tasks as are already performed by
multipart/signed is something to be avoided at almost any cost.
Machines are cheap, but space in the brains of programmers is
not. Creating yet another gratuitous standard is a heinously bad
idea.
> However, I do not feel we should make transports parse MIME at all. I don't
> think the code body is that large.
In terms of raw lines of code, you're right.
In terms of the amount of code out there that parses MIME, well
they're just about the only thing going. Everything else that
attempts to parse MIME almost invariably "borrows" liberally from the
currently existing mail transport code.
> How much code is that? Does any of it relate to the things we wish to
> sign about USENET?
That's pretty much all the MIME parsing code that exists, at
least to the best of my knowledge.
If you want to talk about signatures on messages, and we're
supporting MIME anyway, I still haven't heard anything remotely
resembling a strong argument for why we shouldn't adopt the same MIME
multipart/signed standard that exists today.
Of course, I've said it before, and I'll say it again -- I think
the discussion of signing standard should be tabled for now.
If you're just using this as a bludgeon to get me to agree that
the non-local/non-variant headers should be static and no one should
be able to move them around or otherwise change their formatting in
any way, you've long since done that.
In fact, further discussion from you or anyone else on this topic
is likely to cause me to re-re-consider my position on this issue,
and I might start taking a contrarian view just because you're
starting to really annoy me.
-- These are my opinions -- not to be taken as official Skynet policy ____________________________________________________________________ |o| Brad Knowles, <blk@skynet.be> Belgacom Skynet NV/SA |o| |o| Systems Architect, News & FTP Admin Rue Col. Bourg, 124 |o| |o| Phone/Fax: +32-2-706.11.11/12.49 B-1140 Brussels |o| |o| http://www.skynet.be Belgium |o| \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ Unix is like a wigwam -- no Gates, no Windows, and an Apache inside. Unix is very user-friendly. It's just picky who its friends are.