ietf-822
[Top] [All Lists]

Re: X.400 and extra headers (about stripping them)

1991-10-21 17:31:18
Just a contradiction>
RFC-987, the X.400/84 standard, DOES provide a way of mapping unknown
headers that is reversible.
It is put into a new initial bodypart whose first line is
"RFC-822-headers:"
This is of course reversible, but might lead to header sequence
reordering.

I'm aware of this mechanism. It does not count for two reasons:

(1) The headers are now text, and as such they intrude on the user as text, not
    as headers.  This bothers a lot of people. I do this conversion as an
    option in my gateway, and most sites have elected to switch it off. In 
    fact, I know of only one site that uses it and plans to continue using it.
    (This is based on reports I've received from about 20 sites -- more sites
    are running the code but they don't always tell me how they have this set.)

(2) This mechanism blurs the distinction between text and non-text. It is
    then possible for text to become headers (as well as the other way around,
    which always happens when you do this). This can be a pretty major problem,
    when you get right down to it. Part of this burden is borne by the
    implementor, who must take care not to map damaging headers from one of
    these into the message. But even great care is not enough in all cases.

Given that this whole discussion is about how to handle header display
correctly, I submit that mapping them to become part of the text of the message 
is not a solution that handles the display issues in a reaonsable fashion.
Of course, if your X.400 UA is extended to know that text segments beginning 
with the magic cookie "RFC-822-headers:" are actually headers and it should
then handle them accordingly, then you have a solution which works, more or
less. However, I have yet to see an X.400 native UA that does this.

                                        Ned



<Prev in Thread] Current Thread [Next in Thread>
  • Re: X.400 and extra headers (about stripping them), Ned Freed, Postmaster <=