[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

[gnu.misc.discuss,gnu.emacs.gnus] Free software: Packagers vs Developers

I suppose that many of you have seen the message below posted elsewhere,
but I think that the discussion fits perfectly in debian-devel and

This message was preceeded by criticisms to Debian (not to say bashing)
coming from XEmacs developers at the gnu.emacs.gnus newsgroup in a thread
started with Message-ID: <u8aetkpxsp.fsf@tri-yann.crf.canon.fr>.

Rafael Laboissiere 

------- Start of forwarded message -------
From: Per Abrahamsen <abraham@dina.kvl.dk>
Newsgroups: gnu.misc.discuss,gnu.emacs.gnus
Subject: Free software: Packagers vs Developers
Followup-To: gnu.misc.discuss
Date: 01 Jul 1999 19:03:38 +0200
Organization: The Church of Emacs
Message-ID: <rjpv2cpaj9.fsf_-_@feller.dina.kvl.dk>

[ FUT: gnu.misc.discuss ]

The specific issue is that Debian is distributing an alpha version of
Gnus as the default Gnus for XEmacs on their unstable branch.

There are two problems with this:

1. Alpha releases aren't intended for anyone but developers, and those
users determined enough to subscribe to the developers list, and
download and install it themselves.  

While most projects have adopted a policy which makes it easy to join
as an alpha tester, this should _not_ be thought of as a permission
for outsiders to package and distribute the software more widely.  All
that will accomplish is to make developers more wary of making alpha
releases, and maybe go back to a more closed development model.

2. Gnus is a part of both Emacs and XEmacs, and has been tested to
work together with the other Emacs Lisp packages bundled with each
Emacs or XEmacs release.  By replacing one part of XEmacs with
another, they are distributing something different from what the
XEmacs team released.  This means that it hasn't been tested by the
XEmacs team to work together, and worse, bug reports reaching the
XEmacs team no longer correspond to something they have released.

Both problems are generic, people packaging and distributing alpha
software, and people packaging and distributing their own "improved"
versions of free software.  All the major Linux distributions does it,
perhaps to gain a competitive advantage, but Debian seems to be worse
than the rest.  Perhaps because they have so many package maintainers
(500+), each of whom feel the need to make a difference.

The responses we got on the gnus group from the packagers was:

a) It is legal, so you have no right to complain.
b) _We_ test the code, so _we_ know when it is ready.
c) Our users demands it.

None of which really helps solve the problem.

I know that we developers have to live with _some_ changes being made
in order for the software to work together in a distribution, but it
would be nice if some ethical guidelines among packagers could be
developed to supplement the current Debian rules, like:

1) _Never_ distribute alpha releases.
2) _Never_ distribute improved versions.
3) _Never_ distribute with different "user preference" options 
   than the default.
4) Keep bug fixes and interoperability changes to an absolute minimum.

While this may make the individual distribution "worse" on the short
term, having the code released by the developers being as close as
possible to the code used by the users should improve the quality of
the feedback from users to developers, and improve the quality of the
code for everybody on a slightly longer term.

On the long term, I could see developers themselves create packages
for the hypothetical dominating platform (say LSB) using the
hypothetical dominating package format.  With the right software and
the right standards, it would just be another Makefile target.
Distributors would then collect the already made packages, and we
would avoid the middlemen in most cases.
------- End of forwarded message -------

Reply to: