RE: Need help on AS2

Comments in-line.

> -----Original Message-----
> From: Dick Brooks [mailto:[email protected]]
> Sent: Wednesday, January 10, 2001 11:25 AM
> To: Moberg, Dale; Santanu De
> Cc: [email protected]
> Subject: RE: Need help on AS2
> 
> 
> Dale,
> 
> > Sorry, Dick, but I think this remark is misleading.
> >
> > The so-called "email" packaging makes use of IETF standards
> > for MIME, the security multiparts, and the specifications
> > for Open-PGP and SMIME. It is as much
> > based on standards as is the packaging that makes
> > use of "multipart/form-data."
> 
> My comment about the email packaging of AS2 being 
> non-standard refers to the
> fact that
> neither RFC 822 nor RFC 2045 nor RFC 2616 define the AS2-to 
> or AS2-From
> headers. In fact, these don't exist in the current AS2 draft (ref:
> http://www.ietf.org/internet-drafts/draft-ietf-ediint-as2-07.t
xt) nor any
previous draft of AS2. I know AS2 will be enhanced to support AS2-To and
AS2-From, but as of right now, these headers aren't defined in any RFC or
Internet Draft. Is this factual?


That detail is not a significant basis for saying that 
the _packaging_ is nonstandard. As you well know, HTTP
allows headers to be added without the X- qualifier.
And if we use AS2-To and AS2-From in the final draft,
we can ask IANA to register these, in my opinion.

Some on the list have proposed using different header field
names from AS2-to and AS2-From. 

We should continue to discuss what field to use  on this
list if there are reasons for one way of proceeding rather
than another. And we should discuss these final issues
at an IETF meeting to see if we can determine the consensus.


>I also wanted to point out that the email packaging specification within
AS2
>doesn't follow MIME (RFC 2045) or RFC 822 (which defines the standard
syntax
>for e-mail messages) specifications for extension headers. According to
>these specs extension headers are supposed to contain an "X-" prefix,
>reference this excerpt from section 4.1 (Message Specification/Syntax) of
>RFC 822:


In the same spirit, HTTP doesn't follow every rule of RFC 2045 or
RFC 822. That is one very good reason to stop calling AS2 "email"
packaging. (As you will remember me saying more than once before ;-).
That is one very specific way in which you are being misleading.

We are instead following MIME as used in HTTP which,
as we are at pains to point out in the AS2 spec, differs in some
details from the way it is used in the SMTP context.
Because these differences do not make HTTP RFCs 2068 or 2616
nonstandard, I think it is misleading to say that the use of
MIME in AS2 is nonstandard. Rather the usage of MIME that we recommend
follows how MIME is used in accordance with the standards in 2068
and 2616, and that makes it standard. We are, after all, describing
"HTTP Transport for Secure EDI." and that is why we adhere
to RFC 2068/2616 on those points where HTTP MIME differs from
2045... It might have been simpler if the HTTP and SMTP MIME
were identical (and they are very similar) but that is not
the case. So as a group chartered to find an applicability
statement of existing standards, we have followed the existing
standards even when those standards diverge slightly. (It
is, after all, to be expected that a binary clean transport
might need to say slightly different things about packaging
than one whose default is 7 bit...)