Path: utzoo!attcan!uunet!husc6!cs.utexas.edu!tut.cis.ohio-state.edu!att!chinet!ignatz
From: ignatz@chinet.chi.il.us (Dave Ihnat)
Newsgroups: comp.mail.elm
Subject: Re: New Elm behavior
Summary: The problem is a bugfix introduced a feature change
Message-ID: <8606@chinet.chi.il.us>
Date: 1 Jun 89 18:36:09 GMT
References: <133@dsinc.DSI.COM> <465@wubios.wustl.edu> <5419@pbhyf.PacBell.COM> <469@wubios.wustl.edu> <5423@pbhyf.PacBell.COM> <220@cdin-1.UUCP>
Reply-To: ignatz@chinet.chi.il.us (Dave Ihnat)
Organization: Chinet - Public Access Unix
Lines: 26

If I might make an observation, the problem that seems to have been encountered
isn't even necessarily because there was a feature change.  It is because, once
a version of a program is released, users commonly don't expect the intended
behavior of the release version to be changed by what they expected to be a
simple bugfix.  Normally, such a behavior change would be analyzed, discussed,
and then designed into the next release of the system after developers and
users had hashed out exactly how they wanted the system to behave.  Patch 8
unilaterally and without announcement or discussion changed a very visible
aspect of Elm behavior; this confuses naieve users, worries administrators
(who have to suspect, initially, that they did something to the system), and
can undermine the confidence of the user community in the stability of the
tool.

The cat's out of the bag for this particular patch and function; I'd suggest
a followup patch that makes the new behavior optional, and clean up the
situation for good in the next release.  But in the future, when writing a
patch, perhaps we should all ask "Is this a correction to make the program
honor the documented features of this release, or is it a feature enhancement
and/or functional change?  If the latter, is it important enough to introduce
in the current release, or does it belong in the next release?"  And finally,
if it is important enough to do this, it should be announced *and discussed*
before its posting.

	Dave Ihnat
	Analysts International Corp
	aicchi!ignatz || ignatz@chinet.chi.il.us || ignatz@homebru.chi.il.us