Errors-to: owner-drums@CS.UTK.EDU
Message-Id: <199508241457.KAA09907@wilma.cs.utk.edu>
From: Keith Moore <moore@CS.UTK.EDU>
To: Eric Thomas <ERIC@SEARN.SUNET.SE>
cc: drums@CS.UTK.EDU, moore@CS.UTK.EDU
Subject: Re: Getting back on track 
In-reply-to: Your message of "Thu, 24 Aug 1995 12:28:24 +0200."
             <199508241103.HAA06519@CS.UTK.EDU> 
Date: Thu, 24 Aug 1995 10:57:14 -0400
Sender: moore@CS.UTK.EDU

> I'm afraid this procedure sounds tedious and bureaucratic to me. 

Indeed it does to me too.  But the current procedure is not working.

> It also assumes people are always available to carry out step X
> within Y days, which in the real world isn't the case. And if you
> allow 2 weeks for each step, it will take forever, whereas we have
> limited time.

Okay, I'll note that X and Y aren't fixed constants, that it is
possible to pipeline operations, and that the whole idea of this
procedure is to help us make progress in finite time.  In particular,
it gives us a chance to decide that "this issue's not worth arguing
about" before spending lots of time arguing about it.

It also occurred to me that instead of imposing it on every
discussion, the Chair could invoke it when it seemed to be necessary.

> Finally, there's a kind of assumption that all major user
> groups are (roughly) equally represented on this list, which simply
> isn't true. 

No, that's not an assumption.  We hope that most important
constituencies are represented in a working group, but we don't expect
that they're represented in proportion to their user base.  That's why
we make decisions by rough concensus rather than by simple majority.

> The majority of today's users, who are non-technical and
> aren't interested in arguments of the type "the users are just going
> to have to change their habits and it won't be all that bad", aren't
> represented at all. The recent discussions have been going nowhere
> because it's been a bunch of techies trying to make decisions on
> behalf of users who, whether we like it or not, are the way they are
> and won't change just because a group of techies decided they
> should.

We have the burden of making changes that users either aren't
interested in or don't understand.  Users don't see any reasons why
they shouldn't send 8-bit text mail, unencoded, without a charset
label (after all, *everybody* uses iso 8859/1).  Users don't
understand why you can't reply to their messages when their return
address is joeuser@flathost.  Users don't understand why their mailer
shouldn't be bouncing mail to the header From address.

I'll agree with you that we will have difficulty enforcing a change
that goes strongly against user habits, and we need to keep that in
mind when we propose changes.  We need to assume that any change that
we propose will not be uniformly adhered to, and consider the likely
effects of *that*.  But the argument that "we shouldn't make changes
that users won't understand or that users won't like" doesn't hold
water.

> Since I suspect this is all directed at the Reply-To: issue, let me
> make sure I made my point clear. Users have been enjoying this usage
> of the Reply-To: field for 9 years.  Whether we like it or not, they
> simply wouldn't understand how a bunch of arrogant techies can
> decide that this convenient feature must be removed from them,
> especially without any usable and already available replacement. 

Offhand, I don't recall anyone proposing that the feature be "removed"
or "prohibited". "Discouraged", perhaps, especially when the sender
has already specified a Reply-to field.  But I'm personally on record
as saying that we are probably better off allowing lists to continue
to use Reply-To for that purpose than we would be to create a new
mechanism.

> So, this is a fundamental problem that isn't going to go
> away by making the list more formal or adopting a voting procedure
> or whatever. The bottom line is that no matter what process you use
> to reach this decision and whose signature(s) you get at the bottom
> of the page, if you invalidate the current Reply-To: usage by
> mailing lists I can guarantee that the major players will ignore
> your requirement, because people like to get a paycheck at the end
> of the month. 

You've made this point before.  Please consider the possibility that
you were not ignored when you did so.  Others in this working group
might still disagree with you. Or they might feel that the changes
proposed here are slight enough that they won't have the kind of
severely negative effect that you fear.

> As for the argument about users not being able to understand and the
> chaos we'd have if we let them run the network, that's something you
> can say after 3 months, maybe a year. When something has been
> allowed for 9 years and is still widely used (and showing no signs
> of decline, quite to the contrary), you're stuck with it -
> period. 

Actually, I can think of several uses of protocols on the Internet
which were quite prevalent for years, but which are now being actively
discouraged as the Internet runs into scaling problems.  So I don't
buy this argument either.

In the case of Internet email we have several constituencies that
traditionally have dissimilar behavior (having grown up on BITNET or
UUCP or DECnet or one of the commercial services) who used to be
poorly connected, but who are either better connected now.  We also
have significant numbers of users who have migrated from VAX or large
IBM systems (with their peculiar user agents) to UNIX, Mac, or Windows
systems (and their peculiar user agents), or to one of the many
different proprietary mail systems (which might still be gatewayed to
the Internet).  All of these groups suffer pain when their
expectations are challenged as the result of changing their email
environment, and it's quite possible that these users will understand
the need for a standard that attempts to promote more uniform
behavior.

The Internet has been anything but static for the last 9 years.
There's no reason to assume that just because a behavior has been
allowed for that long, and even that users are accustomed to it, that
it's not time for a change.

Keith