>> As to the ERROR-IS codes, they don't have to be complicated.  When an
>> error condition is detected, they would be an OPTIONAL way to further
>> describe the error.  The only requirement would be that no fales
>> descriptions be sent.  The coding can be structured with varying amounts
>> of information.  Whatever follows the number can be text taken as a
>> comment.
>
>What will programs do with this information about the cause of the
>error?  It seems to me that the only reason for putting such
>information into a header is that a program might process it in some
>way.  A person can see what the error was by reading the text of the
>rejection notice.
 
Well, if there were some clear indication (a separate digit, or dedicated
values in a digit) that indicated a condition that is likely to be
temporary (systems disk space full, gateway queue full, gateway unavailable)
versus a condition for which no relief was in sight (addressee unknown,
restricted gateway), then a suitable program might schedule a retry for
the temporary cases.