[Top] [All Lists]

Re: 7/8-bit conversion vs. bouncing

1991-07-03 18:27:08
most of your scenario disappears if there is an a-priori way to tell if
the remote is supposed to handle "new protocol" (and that you can therefore
wait for the service to be reestablished).  
   But there really isn't other than trying.  Please generalize from the 
reasons why HR banned deciding that someone would, or would not, support 
a particular protocol based on WKS records.

magic cookie information, a ping of some kind, 
   Ignoring "fiat" for the moment, the above are instances of "trying" 
and trigger the problem Craig has identified.

Not a big problem, as long as it is cheap to do offline.
  But you really can't do it offline, you need some sort of failure 
strategy associated with experimenting.  And, if it is two protocols, 
the failure strategy turns into a failure matrix and gets too hard, as 
Craig points out.

Mumble... if I were a dictator (and the IAB is...), I'd just decree
that henceforth all 821 and 822 implementations MUST be 8-bit transparent.
   You think so?  What do you call a dictator that is, in practice, 
totally lacking in enforcement powers and whom people feel free to 
ignore?  I'm especially interested in answers to this question from 
people who will first go back and read the DNS implementation schedule, 
examine the "absolutely final, everyone needs to be running it except 
MILNET sites" deadline and who then is willing to contribute a nickel to 
the charity of my choice for every host-month of non-conformance since 
    This issue has been kicked to death, and I don't want to restart it, 
but to summarize some of the high points...
  o it is morally, politically, and technically unacceptable to declare 
existing conforming systems broken, at least without a requirement on 
the same scale as the DNS.  And, even then, one needs to allow years and 
assume that, in practice, some hosts will take A LOT longer.
  o it would be slightly more plausible if it were true that
the only effort is that a bunch of people need to rip
out the code that prohibits 8th-bit-set characters.
But it isn't.  We have valid, conforming, SMTP servers out there that 
get into deep trouble when someone hits them with eighth-bit-on 
characters.  If they have already gone belly-up, they are not in much of 
a position to "translate".
  The idea is a non-starter, really.