See also: last week's Letters page.

Letters to the editor

From:	 Leandro Guimarães Faria Corsetti Dutra <leandrod@mac.com>
To:	 letters@lwn.net
Subject: GNU-Darwin for the x86
Date:	 Wed, 21 Nov 2001 13:16:19 -0200

"The battle to rename it GNU/Linux has not gone all that far, and 
resentment remains. The same spirit that causes FSF developers to push 
forward with HURD development also draws their attention to other, 
non-Linux alternatives."

	Attribution of motives and sentiments is never good journalism.

	The FSF and the GNU project supporters feels that calling GNU/Linux just 
Linux does a disservice to the philosophical struggle for free software, 
and it is hoped that the Hurd will be a superior kernel to the GNU system.

	But never in RMS writings or in anything published by the FSF or GNU 
project developers I've seen resentment towards whomever call GNU/Linux 
just Linux.  And the naming issue was never the reason behind the 
development of the Hurd; instead, it is believed that the microkernel 
with multiple servers architecture of the Hurd will make for a more 
flexible kernel for developers, testers and users, enabling the Hurd to 
progress more quickly and orderly than the Linux kernel after the Hurd 
reaches critical mass and a stable release.

	Please please please don't put words in other people's mouths.



-- 
   _
  / \  Leandro Guimarães Faria Corsetti Dutra    +55 (11) 5685 2219
  \ /  http://homepage.mac.com./leandrod/        +55 (11) 9406 7191
   X   Orange Telecom                            +55 (43) 322 89 71
  / \  Fita ASCII contra correio eletrônico HTML             BRASIL

From:	 Mark Bainter <mark-spamx@firinn.org>
To:	 letters@lwn.net
Subject: Editorials
Date:	 Mon, 26 Nov 2001 00:04:23 -0600

I appreciated the editorial regarding sourceforge.  Particularly
since it brings to voice concerns I've had privately of late.  Not
only for sourceforge, but also for freshmeat.  Both of which are
extremely valuable resources to the open source community.  Having
not only one in a precarious position like that, but both worries
me and smacks of poor design.  But as you said, it's not like there
is a line of companies looking to relieve valinux of the burden
they have so gladly (or so it seems to me) born for us.  
That said, I appreciate all VALinux has done, and is doing
for us.  I hope I don't sound like an ingrate.  ;-)  Here's
hoping another company steps up to help provide resources to
keep sourceforge/freshmeat alive.  If we could make their 
survival independant of any one companies existance I know 
I'd feel a lot better.

However, the Darwin editorial, while interesting and informative
at first devolved rather quickly into another annoying whine about
GNU/linux.  This is a topic that I (and I think many others, though
I can only speak for me) am really sick of hearing about.  I don't
see the people who make the tools used to build cars lining up to
whine about their names not being on the cars built with them.  I 
don't see lumber companies complaining because the companies building
houses don't include the name of the lumber mill in the name of the 
subdivision being built.  

Linux is it's own product.  The fact that GNU tools are used to 
build it, or were used to write it or (the more common argument)
are used to build a complete OS generally called "Linux" is really
not relevant.  I mean, if solaris suddenly started shipping gnu tools
as part of Solaris instead of their own would everyone be clamoring
to have Solaris renamed to GNU/Solaris?  I'm not a big BSD user,
but don't at least some of the *BSD distros use gnu tools?  Is no-one
going to complain that it should be renamed GNU/BSD?  Doubtfull.  

Come on.  Most everyone in the linux world knows who GNU is.  It's 
all over the place here, and I think most people do truly appreciate
the contributions the GNU foundation has made, and is making.  Can't
we move on?  Hasn't this horse endured enough abuse?  

It's dead Jim, stop beating it already.

From:	 Chris Lawrence <lawrencc@debian.org>
To:	 letters@lwn.net
Subject: bug reporting in noncommercial software
Date:	 Thu, 22 Nov 2001 01:52:12 -0600
Cc:	 Mark Bainter <mark@firinn.org>, Seth LaForge <sethml@ofb.net>,
	 David.Kastrup@t-online.de

As the author of Debian's reportbug, I'd like to thank Seth for his
comments about reportbug in the 11/15 LWN.  I'd also like to respond
to Mark's comments about bug reporting in general.

Mark says that each package should include its own bug reporting
frontend.  While laudable in theory, this introduces a number of
problems:

1. Lots of additional bug reporting tools.  On my Debian box, I have
bug reporting tools for mutt, libc, KDE, GNOME, and a few other
programs.  If everyone did it, we'd have a veritable raft of bug
reporting tools.

2. Lots of version skew.  In the case of mutt, its bug reporting tool
is a fork of Debian's "bug" command (reportbug's older, sh-script
sibling).  If we find bugs in "bug", their fixes have to be propogated
over to mutt's tool.

3. No real prescreening of bug reports.  Most people get their
software through distributions.  Probably 1/3 to 1/2 of problems
people have with software are distribution-specific issues (why didn't
X pull in library Y when I installed it, etc.).  If the libc people
say "report all libc bugs using glibcbug", they'll get a large number
of reports that are Debian's or Red Hat's or SuSE's fault.

Mark does raise a valid issue about what sort of information should be
included in bug reports.  A standard reportbug report includes:

- The package and version
- The specific file mentioned by the submitter (if specified)
- A severity tag used by the BTS and maintainers for triage
- The body text written by the user
- The output of uname -a and a few locale settings (LANG, LC_CTYPE)
- The first-level dependencies of the package, with versions
- Any modified configuration files (optional)

However, reportbug (and bug) include hooks for allowing them to report
additional information about the package.  Not many packages take
advantage of this, however (perhaps because it's poorly-publicized).
This allows what Mark wants: package-specific data collection, or even
an interactive troubleshooter.  For example, a picture viewer might
include a bug script requesting that the user try different X or
framebuffer settings, or identify whether the problem only affects
certain image formats.

reportbug also includes hooks supporting submission to different types
of bug tracking system; GNATS support was added for the now-no-longer
Progeny distribution, for example.

I'll be the first to admit reportbug isn't perfect... it really is
newbie unfriendly in places, for example, something I'd like to work
on.  But it definitely is something a "universal" free reporting tool
could be based on, and I'd be happy to add code to separate it from
its Debian-centricity (easy enough to do, really... just figure out
what distro it's running on and behave accordingly).


Chris
-- 
Chris Lawrence <lawrencc@debian.org> - http://www.lordsutch.com/chris/

From:	 Richard Kay <rich@copsewood.net>
To:	 dave@userland.com
Subject: RMS
Date:	 Sun, 25 Nov 2001 21:08:44 -0500
Cc:	 rms@gnu.org, letters@lwn.net

In http://scriptingnews.userland.com/backissues/2001/11/21 
   dave@userland.com  wrote:
> Do you work really hard to make good software? I do it every day. Does
> Stallman push the envelope? I haven't seen any evidence of that. Imho,
> the economy is still rewarding the wrong people. At one time if you
> pushed for excellence in software, you could build a nice business. I
> still believe that. But it's disheartening to see so much money go to
> support Stallman's theories. I believe this works against software
> breakthroughs, even software progress.

Richard Stallman (RMS) has probably done more than anyone to promote 
software reuse. In this respect his GPL represents one of the most 
effective works of software engineering ever written. Being able to 
modify a world-beating functional system (GNU/Linux) to suit my own 
needs such that I only have to focus on the coding which
modifies it to make it suit my own requirements ( e.g. see 
http://copsewood.net/shared-mailbox/shared-mbox.html ) has
largely been made possible by RMS's ground-breaking work. The
technical achievements of those working on the Linux kernel
and compatible free-software application infrastructure has
also been made possible, in no small part, due to RMS's direct work
earlier on the GNU C compiler and Emacs editor. There have
also been much greater indirect benefits through the
improvements for free sofware brought about by the difference
between BSD style licenses, which allow for the tragedy of
commons arising from the theft and distortion of free software 
by commercial interests who are obliged to give nothing back, 
and the GPL which encourages a more open and community-oriented 
style of software development.

> Something to think about. Would the $830K have been better used to
> support SourceForge?

Probably not. While SourceForge has given practical help to very
many projects, there are many willing to host such projects and
it is probably undesirable to have too many free-software
projects hosted un-mirrored on a single centralised server, subject
to whatever legislation a single nation's government and lobbyists 
might dream up. The potential benefits of promoting free-software
are very likely to outweigh supporting what should become a
profitable and self-supporting Internet business. 

It is possible that the same argument could be used to say
that supporting free software, if it is as useful as I suggest,
could also be carried out on a commercial footing. I have to 
disagree with this counter-argument, as the moral benefits of 
free software probably far outweigh the practical and commercial 
ones, in the sense that free software can act as a potentially 
liberating and democratising influence in areas other than just 
ICT, e.g. in areas as diverse as privacy, integrity of electronic voting 
systems and the ability of musicians, writers and artists to 
bypass corporate distributors who have traditionally controlled most 
intellectual property rights. These benefits should not be
lost and distorted through legislation sponsored by commercial 
interests such as the DMCA and SSSCA.

The papers which I have written and published on http://copsewood.net/ 
which are concerned with the potential for a more democratic,
sustainable and decentralised society are unlikely to have become
possible without the influence which has derived directly from
RMS's work.

Richard Kay
Senior Lecturer/Technologist
Technology Innovation Centre,
University of Central England,
Birmingham, UK.

From:	 Richard Stallman <rms@gnu.org>
To:	 rich@copsewood.net
Subject: Re: RMS
Date:	 Mon, 26 Nov 2001 04:48:45 -0700 (MST)
Cc:	 dave@userland.com, letters@lwn.net

Thank you very much for speaking up in my defense.  I would like
to correct one factual point about the prize itself.

$830k (actually a little less with current exchange rates) is the
total sum.  Since it is being shared by three people, I will get 1/3
of that--after taxes, perhaps $170k.  It's a nice sum of Hanukkah
gelt, and will make a difference for me, but it wouldn't support an
organization like Sourceforge for long.

From:	 "Jay R. Ashworth" <jra@baylink.com>
To:	 letters@lwn.net
Subject: Dave Winer and RMS
Date:	 Sat, 24 Nov 2001 14:03:44 -0500

Well, here we go again...

You know, it's funny.  Dave Winer used to be one of my favorite people.

I was lost in the world of Macintosh, back in 1992 -- I was helping
start a cool TV network called MOR Music TV, now, sadly, defunct --
when Frontier was in about release 3.something, and I fell in love with
it.  Given the lack of a command line on a Mac, Frontier was about the
closest you could get, and I liked the outliner-based approach to the
whole thing.  A lot.  Same reason I like Zope -- or I think I would, if
it would stop moving long enough for me to figure it out.  (Friendly
big Zope site administrators cheerfully solicited...)

Even though it's one of the best examples of "when all you have is a
hammer, everything looks like a nail", it worked, and it worked pretty
well.

And then I actually tried *interacting* with Dave.  I wrote to him, having
become a regular reader of his Scripting News weblog -- it was even my
browser homepage for some time -- and suggested that he might want to
look into Linux, and even maybe doing a port of Frontier to Linux.

He called me every kind of a mother-fscker, and several I think he made
up on the spot.  And then, less than a year later... he fell in love
with Linux.

So it shouldn't be any surprise that I take most of what Dave says with
a grain of salt (even though Doc Searls, whose opinions I respect
highly, doesn't).  But I think that his comments on RMS, quoted in this
week's LWN are particularly off base, and I'll tell you why.

"It's disheartening," Winer says, "to see so much money go to support
Stallman's theories."

I disagree -- although you probably had already figured that out by
now.  I've watched "open source" software for a very long time; I go
back to at least 1982 on Usenet, and a bit before that in working with
Unix, Xenix, and their ilk.  I am right here to tell you that the
driving force behind the creation and expansion of large,
multi-programmer projects in that arena has been Stallman's General
Public License.

I can't see anything else that could have made possible projects like
perl5 and 6, PostGreSQL 7, and, indeed, the Linux kernel itself --
RMS's insecure grumbling about why it's not referred to as "GNU Linux"
notwithstanding...


I also spend a fair amount of time working with the HylaFAX
(http://www.hylafax.org) fax server software package.  Originally written
by SGI's Sam Leffler, and mostly maintained by him up til about it's 4.0
release, the package got sort of stagnant for some time.  It now has 5 or 6
pretty sharp people working on it, and it's moving along again
nicely... but I can't help but wonder: is the reason that it has trouble
attracting even more people motivated enough to work on it that it is *not*
licensed under the GPL, but rather, under a license roughly equivalent to
the BSD license (which doesn't protect potential contributors from
commercial entities making off with their hard work without any recompense,
credit- or otherwise)?


General George C. Marshall, US Army Chief of Staff during WWII, and
author of the "Marshall Plan" -- which helped rebuild Europe after the
war and gained him a Nobel Peace Prize -- is most generally credited
for the observation that "there's no limit to what a man can achieve if
he isn't concerned whether or not he gets the credit for it". 

The GPL is my favorite example of this, with the delightful twist of
irony that it works almost precisely by preserving the credit due to
those people who write the code released under it -- which is all it
preserves.  The only person whose credit isn't really that well
preserved is RMS's.


So, for putting up with 20 or 25 years of the lifestyle engendered by
the beliefs that gave us the GPL and, hence, the OS running on the
laptop I'm writing this letter on, hell yeah, I think RMS is entitled
to the prize he's been awarded.

There's nothing wrong with asceticism... except for that class of
problems that money is all it takes to fix.

---

On a final note, I find it amusing that Winer snipes at Danny O'Brien,
of NTK.  O'Brien is sitting there, in front of Linus and everybody,
asserting that Winer's also done something worth rewarding, and Dave
gets pissy over it.

Some people just think too much, I think.

But who knows; maybe it's just me.

So many things are just me.

Cheers,
-- jra
-- 
Jay R. Ashworth                                                jra@baylink.com
Member of the Technical Staff     Baylink                             RFC 2100
The Suncoast Freenet         The Things I Think
Tampa Bay, Florida        http://baylink.pitas.com             +1 727 804 5015

   "If you don't have a dream; how're you gonna have a dream come true?"
     -- Captain Sensible, The Damned (from South Pacific's "Happy Talk")

From:	 Micah Yoder <micah@yoderdev.com>
To:	 letters@lwn.net
Subject: Future of SourceForge
Date:	 Wed, 28 Nov 2001 00:58:14 -0500

Hi,

There seems to be a sentiment out there that SourceForge is in danger of 
being shut down by VA Linux.  As you said in your 11/22/01 front page,

> SourceForge is an expensive gift from VA Linux to the free software 
> community; if VA continues to bleed cash and continues to move toward 
> proprietary software, the company will eventually be forced to look at 
> ending that gift.

Frankly, I don't think we have anything to worry about, at least for a couple 
years.  SourceForge (the now proprietary software) is the cornerstone of the 
new VA Linux business model.  They are focusing on selling this software for 
use in Global 2000 companies.  But how are they going to inspire confidence 
in those companies and sell their software as a solution?  How are they going 
to prove that their main product is useful and scalable?

Right!  SourceForge.net!

In addition to being proof that their primary product works and 
enterprise-ready, SF.net also ensures that there are thousands (nearly 
300,000 actually) of users who are familiar with their product, based on 
their Open Source work hosted at SF.net.  Many of these users will then, 
supposedly, recommend SourceForge Enterprise to their employers.

The bottom line is that VA Linux cannot possibly afford to take SF.net down.  
It would be suicide!  I am therefore convinced that it will be around as long 
as VA is in business.  (Unless they change their business model again, but we 
won't get into that!)

That brings up another question:  How long will VA be in business?  According 
to their recent annual report, they had $60 million in cash as of July 28.  
Granted, they had $123 million a year prior, but 1) they now have fewer 
employees and 2) last year involved some enormous expenses involving their 
changed business model.  They should be able to last AT LEAST another year 
without making a dime.  But they already have some enterprise customers, and 
hopefully they will get more this year.

So don't worry about SourceForge.net.  It won't be disappearing anytime soon.

Micah

-- 
Like to travel?                        http://TravTalk.org
Micah Yoder Internet Development       http://yoderdev.com

From:	 "Jonathan Day" <jd9812@my-deja.com>
To:	 letters@lwn.net
Subject: The folly of slowing down
Date:	 Wed, 28 Nov 2001 07:59:07 -0800

Dear editors,

  I have to disagree with Linus Torvalds (gasp!) when he argues that there
need to be fewer fundamental changes. The problem is that "stable" code is
also stagnant code, which means that as the dependencies age, it becomes
"unstable, but intractible" code.

  This is one of the genuine problems that is half-jokingly referred to as
"bit-rot". Of course, the bits don't actually decay with time, but the
assumptions on which they are built -do-. And that can kill an OS.

  What we need are far MORE fundamental changes between stable
releases. Keep EVERY element of the kernel alive. If there is a single line
of actual code older than a year in the kernel, then someone is being
slack. Either the maintainer isn't refining their skill set (and thereby
rotting, in themselves), or the code isn't being scrutinised nearly often
enough for potential bugs, security holes, etc.

  Replacing the virtual memory system took over 170 patches, if I
understand correctly. Far too many. A sign that the code isn't being
actively worked on, nearly enough. Why? Because no sane coder would develop
code that hard to maintain, if they were actively thinking about it. You
just can't keep track of 170 fragments of code as easily as you can one
self-contained unit.

  IMHO, 2.5.x needs one gigantic, fundamental change, if it is to survive
another 10 years.  It must be ripped apart, and sewn together, as many
times as it takes to seperate out entangled code. (The IPv4/IPv6/IGMP
entanglement is positively horrible! IPv6 development is now -years- behind
other Linux IPv6 stacks, we STILL don't have IGMPv3, it's not possible to
have an IPv6-only box, IPv6 netfilter can't do a quarter of the things
IPv4's can, and those are just the problems I've noticed.)

   The day Linux is allowed to stagnate is the day Linux will die. I know
this is personifying it slightly, but oh well. Linux lives to grow, and
grows to live. It is, in a sense, a living thing. You feed it and nurture
it, it'll thrive. Cut it down, to "preserve" it, and all you have is a dead
thing.  Is that what we want?  Really?

Jonathan Day


------------------------------------------------------------
--== Sent via Deja.com ==--
http://www.deja.com/