Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!cbatt!osu-eddie!lien
From: lien@osu-eddie.UUCP
Newsgroups: comp.unix.questions
Subject: withdrawing E-mail
Message-ID: <3152@osu-eddie.UUCP>
Date: Wed, 18-Feb-87 23:42:50 EST
Article-I.D.: osu-eddi.3152
Posted: Wed Feb 18 23:42:50 1987
Date-Received: Fri, 20-Feb-87 00:54:37 EST
Sender: lien@osu-eddie.UUCP
Organization: OSU
Lines: 61

From: Yao-Nan Lien <lien>





Thanks for all who tries to help me on this issue. 
From all the responses we have seen,  we can conclude that
a user should be more careful since there is no 'undo' feature
for E-mail once a '^D' is typed.

As a user, I believe  that this  kind of mistake  may be inevitable. 
Human being makes error, isn't it?
I believe it is important and serious for the following reasons:

1. E-mail is getting more important than before in a computer system.
   Take me as a case. Writing paper, E-mail, USENET communication
   are my major computer activity. This may not be true for
   super users, but it may be true for many regular users.

2. It is serious since no 'undo' function is provided.

   Some people may say that 'rm *' is more serious than making
   mistake in a E-mail. But, we have backup system to reduce its
   damage. People invent backup system but not undo function for
   mail is simply because E-mail wasn't an important function long 
   time ago. Now, things are changing. It  is time for us to
   redefine the desired functions for regular users.

So, I am suggesting to take this into consideration for the
future OS design.

Nothing is perfect. There is no complete solution for this
problem. Let's see how much we can do first.

1. The theoretical latest time that one can withdraw letters
   from his own mailbox  is the time the mailman comes.

   For those who use apartment shared mailbox, this time is much
   earlier. One can only do it before he drop the letters into the
   mailbox.

2. This theoretical latest time on computer system (local or
   remote) is the time right before the receiver read it.


3. With the help of super user, the latest time in current
   UNIX system  is the time before it is sent out to other net
   or the time it is read by the receiver if it is local.
   (Am I right?)

4. Without any help, the latest time is the time one hits '^D'.


In the future system, what is the latest time for a regular user?
Can we try to extend it to the theoretical upper bound?
What is the difficulty?
Is  it worth to do?


Yao-Nan Lien