On Aug 5, 10:01, Eric Thomas wrote:
} Subject: Re: SCIFAQ errors
} On Wed, 4 Aug 1993 19:21:26 -0400 Chris Lewis
} <[log in to unmask]> said:
} >No, I'm applying usenet logic to usenet. Articles posted to usenet
} >should never bounce. I don't care what a gateway is gating to,
} >transmission problems outside of usenet are the gatewayer's problem.
} Correct. It's magic wand logic you're applying here. You are posting to
} usenet and expect a certain behaviour, which is standard in usenet.
} Someone is gatewaying your message to another world, which expects a
} different behaviour, standard in that world. You expect the software in
} that other world to somehow divinate that this particular message comes
} from someone who posted it via usenet and who expects a different
} behaviour.
Correct. A proper gateway from one world to another should conform to
one world on one side of the gate, and the other world on the other side
of the gate. This is a surprise? That's what gateways do, right? A
gateway that can't guarantee that isn't a gateway and should be disconnected.
} Furthermore you expect that software to accomodate your
} expectations even though they are the opposite of its normal operational
} mode.
False. I expect the gateway to accomodate my expectations of behaviour
on my side of the gate only. The other side of the gate can do
whatever it wants, but it shouldn't enforce it on my side of the gate.
If the LISTSERV/list-owner want to prohibit duplicates on the LISTSERV
within a certain time period, fine. If it wants to bounce them, fine.
But do not enforce bounces on USENET where they violate the standards.
} Of course the rest of the world should adapt to your expectations.
You've changed the definition of world here from what you were using
before. You mean "worlds" or "universe" now.
False. I do not expect other worlds to adapt to my expectations, or
rather, the standards of my world. But where those other worlds
affect my world, they'd better conform to the standards of my world.
} If there is a possibility that you might receive a message you don't
} want, the software that sends it should be modified to never send it.
Correct. The standards of my world, my side of the gateway, forbid the
gateway sending bounces to my world.
} After all, since it is never useful to you, it is never useful to anyone,
This is only one of the alternatives I suggested that would bring the
USENET end of the gateway into compliance with the standards of USENET.
Another alternative is that the LISTSERV gateway divine the origin of
the messages, and behave appropriately (in this case, silent discard
with/without logging).
Explain to me again, how a gateway that is explicitly retrieving USENET
articles cannot tell that the articles it retrieves are from USENET.
Explain to me also how a LISTSERV gateway that is explicitly designed to
to carry USENET FAQs can't tell that it's carrying USENET FAQs.
} and everyone should follow the usenet design.
False.
} God forbid that there might be different philosophies, cultures or designs
} in the network.
God help us when BITNET philosophy, cultures and designs determines how
USENET works. Next thing you know, USENET-to-LISTSERV gateways will
start reposting USENET article back into USENET in uppercase - and you'll
continue arguing purity of essence.
You wouldn't be arguing if LISTSERV bounced articles back to USENET
posters if the USENET articles exceeded a LISTSERV size limit. So
why here?
The USENET-to-LISTSERV gateway should conform to USENET standards on
the USENET side of the gateway. That means that bounces to USENET
articles should be prohibited. I don't care how you do it - I've
merely made several suggestions on how you accomplish that. If you
still can't stop the bounces, still can't provide any mechanism
that allows the list-owner to selectively stop the bounces, then
disconnect the gateway.
--
Chris Lewis; [log in to unmask]; Phone: Canada 613 832-0541
Psroff 3.0 info: [log in to unmask]
Ferret list: [log in to unmask]
|