Re: [VPIM] Re: IETF-FAX WG Draft agenda at Yokohama

[Dave Crocker <dcrocker@xxxxxxxxxxxxxxx>]

Graham,

At 08:23 PM 7/9/2002 +0100, Graham Klyne wrote:
> If you were to do this, I think you'd also need to state, reasonably 
> tightly, the conditions under which the esmtp negotiation would be used 
> rather than the UA-based negotiation.
...
> > > At 09:53 PM 7/7/02 -0700, Dave Crocker wrote:
> > Toyoda-san and I suggest that the requirement for conforming to 
> > fax-oriented use of esmtp-conneg, such as with citations to:
> > ...
> > be added to FFPIM.


I agree with your concerns.  So let me wander through some initial thoughts 
on the matter:

1.  I consider FFPIM to be our final effort at achieving a "complete" 
emulation of facsimile through an email profile.  Therefore I suggest that 
we try to incorporate as much as we can, although I acknowledge the 
inherent project management danger in ever making such an open-ended statement.

2.  Both the UA and the SMTP mechanisms use the same CONNEG information 
structure.  So the difference, here, is about the mechanism for exchanging 
things.  (I am being simplistic, but I think it reasonable for this level 
of discussion.)

3.  Given that we are standardizing two ways of achieving the same content 
tailoring -- or rather, two ways of carrying the same content tailoring 
guidance -- I believe it can only be a Good Thing to try to guide and 
resolve the choice within a single document.  In fact, we might even get 
some integration.

So the first thing we need to do is to specify the fax-specific content 
details.  This is the detail in the two RFCs.  It should be the same 
capabilities information for both mechanisms.

My immediate thought is to couple together the use of the two 
mechanisms.  Try to use the SMTP mechanism and fall-back to the UA 
mechanism, because the SMTP mechanism is likely to work more 
quickly.  (These are all MAY, rather than SHOULD or MUST 
specifications.  The goal is to specify a means of integrating them, rather 
than to mandate the One True Way.)

So, we list each mechanism.  Then we list a way to use them in a 
complementary fashion.  (And we might even make a non-normative comment 
about other possible means, such as a hypothetical LDAP-based query.)

Perhaps we need to make a statement about precedence for using information 
obtained through different means.  Come to think of it, we probably need to 
do that anyhow.

OK.  That's enough theorizing, for starters.

Reactions?

d/

----------
Dave Crocker  <mailto:dcrocker@xxxxxxxxxxxxxxx>
Brandenburg InternetWorking  <http://www.brandenburg.com>
tel +1.408.246.8253;  fax +1.408.850.1850