Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!harpo!seismo!hao!hplabs!sri-unix!goldfarb.ucf-cs@rand-relay
From: goldfarb.ucf-cs@rand-relay@sri-unix.UUCP
Newsgroups: net.micro.cpm
Subject: Re:  mdm711.asm
Message-ID: <3681@sri-arpa.UUCP>
Date: Sun, 7-Aug-83 01:15:43 EDT
Article-I.D.: sri-arpa.3681
Posted: Sun Aug  7 01:15:43 1983
Date-Received: Tue, 2-Aug-83 11:53:31 EDT
Lines: 96


<<>>

Keith,
	I am once again incensed enough about the MODEM7 distributionitis
issue to revive the horse we beat to death a few months back.  My ire has
nothing to do with your correspondence with Russ; it is a reaction to your
announcement of the impending release of MDM712 a mere 29 days after MDM711
was released.  Since you have accepted the role of coordinator of RCP/M 
and public-domain software to a certain extent, I once again appeal to you
to help cause more control to be exercised over future releases of MODEM7.

	Here is a little history of the most recent releases along with 
the intervals between them:

              MDM712    7/30/83 
				 29 days
              MDM711    7/01/83 
				  9 days
              MDM710    6/22/83  
				 26 days
              MDM709    5/27/83  
				 12 days
              MDM708    5/15/83  
                                 34 days
	      MDM707    4/11/83
				  7 days
              MDM706    4/04/83

	My original appeal called for a moratorium on new features and a 
quarterly release schedule with intervening bug fixes in the form of diffs 
or DDT instructions.  I, too, maintain an RCP/M system here which, while not
as active or widely-known as yours, still has "customers" who are
knowledgeable and who want the lastest version of everything.  I don't 
blame them for that so I try to keep current.  But sometimes I will no
sooner have one version of MODEM7 on the system than another is released.
This gets old rapidly.  Not only do I have to procure and upload the basic
files such as MDM711.ASM (yes, Keith, the source -- we can cry about not
needing it until we're blue in the face, but people still want it),
MDM711.COM, etc., but also I must do the same with ALL the overlay files
which are changed in a miniscule fashion each time the package is upgraded.
There were 21 of these when I last looked. 

	Granted, the new features have been attractive and desirable.  But
in some cases there has been insufficient time for even a decent alpha
test.  The bugs are frequetly caught during beta testing which is done by the
end users (contrary to the notion of what a beta test should really be).
If Berkeley managed bsd Unix like MODEM7 is managed, I'd have to install
and debug a new operating sytem here every month or so.  (OK, perhaps this
is an unfair comparison -- maybe we should look at Rick Conn's excellent 
management of the ZCPR2 project as a better benchmark.)

	Let me restate the points I'd like to see addressed by this group:
            1)  A suitable time frame for major releases.  If we could 
                cause releases to be made no more frequently than quarterly
                it would allow time for sufficient alpha and beta testing.
		It would also ease the burden on RCP/M sysops.
            2)  Interim bug fixes and *minor* enhancements through the use
                of diffs and DDT patches.  If the program is well tested
                and features are well thought out, there shouldn't be too
                many of these.  
            3)  Generalization of the main program and the configuration
                modules such that each overlay file does not have to be
                regenerated for every new release of MODEM7.  As I
		mentioned above, there are now 21 overlays for various
		hardware configurations.  This number will surely grow
		as time goes on.
            4)  An end to the discussion about whether the source for 
		MODEM7 should be distributed.  I believe it has been shown
		that users do want it.

<<>>

	Yes, I realize that we're dealing with public domain software that
represents the fruits of Irv Hoff's unpaid avocational efforts for which
he is seldom thanked.  Don't get me wrong, Irv has done a TREMENDOUS job
and he deserved to be thanked profusely for his labors.  I believe he has
finally refined MODEM7 to the point where it is robust enough and 
sufficiently generalized to make it an important part of every CP/M user's
software library.  Indeed, with the release of MDM711 I have put my money 
where my mouth is, junking MODEM216 and finally joining the rest of the world.
Also, I am not scorning your efforts on behalf of the CP/M community.  Your
tireless work is infrequently recognized yet you continue to spend much
time to keep the rest of us happy.  If it seems you are the target of this
flame, I apologize for that.  I merely want to see if we can jointly (all
of us) get a handle on the release situation and approach future releases
with a modicum of restraint and control.  I believe this will lighten the
load on Irv, yourself, and the rest of the MODEM7 distributors and users.

				Ben Goldfarb
				ARPA: goldfarb.ucf-cs@Rand-Relay
				uucp: {decvax,duke}!ucf-cs!goldfarb

P.S.
I'd appreciate it if someone could post this on CompuServe where Irv can
see it.