[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Guaranteed 8-bit path to conversion machine?
- To: ietf-822@xxxxxxxxxxxxxxxxxx
- Subject: Guaranteed 8-bit path to conversion machine?
- From: avi@xxxxxxxxxxxxxxx
- Date: Thu May 30 22:31:53 EDT 1991
- >to: ietf-822@dimacs.rutgers.edu
- Email-version: 2
- In-reply-to: your message of Wed May 29 10:40:32 -0400 1991
- Message-version: 2
- Original-from: avi
- Phone: 908-576-3218
- Ua-content-id: <PMX-TERM-2.2-999981-pegasus1-avi-4302>
- Ua-message-id: <PMX-TERM-2.2-pegasus1-avi-2982>
I have a question about the proposals to have an 8-bit message routed to an
agent that knows how to convert it to an appropriate 7-bit version. I am
glossing over the subtle distictions whether the user can specify one or more
machines/agents/programs to do the conversion.
If I understood correctly, the user on machine A is sending a message through
a series of hops (passing through B, C and D) on the way to E. If any
intermediate machine (such as C) is not able to pass it on to the next step,
because of the 8-bit problem, then it is requested to route the message to
machine Z -- which is known to be able to do the conversion. I see one problem
with this. Machine Z may not be easy to connect to from C. Perhaps the routing
is through X and Y. If either of these machines share the 8-bit problem, how
should this be handled? Yes, there may be alternate machines, but the
alternates may also need to be routed through such blind loops.
I would be more comfortable if all places where a conversion needed to be made
could do the conversions themselves following a clear set of rules about what
conversions should be done.
Avi E. Gross @ AT&T LZ 3B-211 (201) 576-3218
attmail!avigross or att!pegasus!avi
avigross@attmail.com pegasus!avi@att.com