Re: FSP-1040.001 draft 1
  By: mark lewis to Stephen Hurd on Fri Mar 04 2016 08:49 pm

 >  SH> This spec only covers the packet formats.  Any cost field in some
 >  SH> other file or structure isn't specified here.  I could remove the
 >  SH> "Usually set to zero" part if you like, though I think having it in
 >  SH> there as guidance is worthwhile.
 >
 > no... it is fine there... i was thinking that an explanation for why it is
 > set  to zero would be helpful...

There is no documentation I could find on how to set the cost value of a
packet... and the individual messages in the packet don't have a cost.  There's
not even any documentation of what the cost represents.

I could add that it's the expected cost the sender bears for delivery of the
packet, but it's not clear if it's sender costs, receiver costs, or the sum of
both.  Without any clear reference for what that value should be, I can't write
useful documentation on it... I mostly just copied the FTS-0001 information
there.

 > if you and i are linked, if my system places a value in the cost field of a
 > packed message, what is your system supposed to do with that cost value when
 > your system unpacks that message for processing?

Nobody knows because the meaning of that value was never documented.

 > zero it on  reception if it is not already zero... the receiving system may
 > alter the value to fit their cost structure when they quality the mail for
 > the next hop (eg: routed netmail being processed on an intermediate
 > system)...

Except not all packed messages will necessarily have the same next hop.  It
could change that to the cost that was borne on receiving the packet, but I'm
not certain how useful that information would be.
--- SBBSecho 2.32-FreeBSD