Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site sneaky.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxj!ihnp4!inuxc!pur-ee!uiucdcs!sneaky!gordon
From: gordon@sneaky.UUCP
Newsgroups: net.notes
Subject: Re: Orphaned Response
Message-ID: <-1087519@sneaky.UUCP>
Date: Fri, 19-Oct-84 03:10:00 EDT
Article-I.D.: sneaky.-1087519
Posted: Fri Oct 19 03:10:00 1984
Date-Received: Tue, 23-Oct-84 01:29:31 EDT
References: <1679@inmet.UUCP>
Lines: 33
Nf-ID: #R:inmet:-167900:sneaky:-1087519:37777777600:1125
Nf-From: sneaky!gordon    Oct 19 02:10:00 1984

Even if remote cancellation is not allowed, you can still get into
trouble if anyone expects references like "the previous response",
"note 7 in net.garbage", etc. to mean anything.  Example:

	A writes a note.
	B writes an idiotic response.
	C and D start writing responses complaining about B's response,
	but referring to it only as "the previous response".  

	Whichever of C or D finishes his note last will probably look like
	an idiot.  

The time windows for such things get even wider if you assume that C and D
are on different machines.  Then it is likely that C and D both look like
idiots on the other's machine, but sensible on their own.

Also, how about:

	A writes a note.
	B writes a nasty response, referring to A's note as "note 7".
	The nightly nfarchive is done, and B's note is now note 7.

	Everyone thinks B is an idiot for fighting with himself, and losing.


Ray, I think I agree with most of your reasons for not having remote
cancellation.  But I think your examples stretched the point a bit too far.


				Gordon Burditt
				convex!ctvax!trsvax!sneaky!gordon
				microsoft!trsvax!sneaky!gordon