I  now  have a  super-duper  RFC822  header  parser, which  supports  nested
comments,   all  forms   of   userid/name   specifications,  comments   within
userid@node,  etc. Problem:  what do  I do  with those  $"!%$!%+ invalid  mail
headers? Here's one:
 
From: "blurb blurb blurb SOMEONE at SOMENODE" <SOMEONE>
 
  I don't  know what kind of  software generated that, but  nevermind. The old
code would accept it --  I don't know what it would do with  it, but I KNOW it
does not reject  the thing. Now my  RFC822 parser says "Mail  origin cannot be
determined", which is  perfectly true and tells the caller  to throw the thing
away. It  probably thought  it to  be SOMEONE@localnode  or suchlike,  but the
thing went through the mail exploder.
 
  The question is, should  I install the parser in LISTSERV  or discard it and
increment my  "useless program (purely  wasted time)" counter?????  The parser
does not  know the RSCS file  origin, I don't think  it should know it,  and I
don't think it  should accept semi-partial portions of an  address as 'valid'.
On the other hand I don't think  LISTSERV should discard something based on an
unknown  mail  origin  --  I'd  get  the mailerless  MVS  people  on  my  back
immediately, not to mention the hacked-IBM-NOTErs.
 
  It was really a good idea to adopt RFC822 as the official internet standard.
 
"Woe to her  chance, and damned her loathed choice!  Accursed the offspring of
so foul a fiend!" <[log in to unmask]> ";-)"
 
  Eric