See also: last week's Back page page.

Linux links of the week

The MetaChart is an accumulation of comparison information between Linux and Windows NT. Not perhaps the most fun everyday reading, but it can be a good source for advocacy material.

Section Editor: Jon Corbet

Letters to the editor

To: editor@lwn.net
Subject: Distribution Inconsistencies FUD
Date: Thu, 04 Mar 99 18:12:59 -0800
From: Dan Benson <bensond@ssds.ucdavis.edu>


I am going to have to disgree with Conrad Sanderson's letter 
to the editor in regards to inconsistencies between GNU/Linux
distributions.  This is indeed a very serious problem.  The LSB
wouldn't have formed if it were not a problem.  I think it is 
an even more serious issue now than ever before.  Let's face it
commercialism in the GNU/Linux community is here.  Companies
have and will port to GNU/Linux.  In fact, many companies
have annouced the various distributions they will be supporting.   
This has created a whole new problem.  If the LSB and it's 
participants are not successful then other "not so big" distributions
will most likely have to conform to the distributions that these
companies are supporting. In my mind this takes away a very large
chunk of free out of freeware.  Sometimes (actually, most of the time)
I think Richard Stallman has it all together and the rest of us 
are nuts.  

Plus, libraries are not the only issue here.  For example, some 
distributions choose to use sysV startup scripts whereas others
prefer bsd based scripts.  This minor detail is not a problem 
for more UNIX types but for a company to support both cannot be 
expected.  A good solution to this would be the way FreeBSD 
handles it, but this is just my opinon.  

Obviously, the LSB isn't the best solution.  The little guy who
wants to be different is always going to get screwed.  But, 
I will say this, I would much rather prefer the LSB making 
the decisions than the most popular GNU/Linux distribution.  

In affect, the Linux community has created not just a single 
UNIX based operating system, but rather several of them.  Even
worse I can't even do something like 'uname -a' to find out what
is being run on a particular machine.

I use GNU/Linux religously, but I am also not blind to the sort 
of problems it faces.  

-Dan

Date: Sat, 27 Feb 1999 14:19:42 +0000
From: Thomas Hudson <thudson@cygnus.com>
To: lwn@lwn.net
Subject: Kudos to Trident

 I'm writing you to let you know of a recent hardware company's
 exceptional support to the GNU/Linux community.

 Trident Microsystems (http://www.tridentmicro.com)
 recently contacted the ALSA developer mailing list to announce that
 they had written ALSA compatible drivers for their 4D Wave PCI sound
 card chipset.  Trident has offered the source for these drivers, as
 well as complete technical documentation for the chipset.  All of
 this material has been provided voluntarily, without external
 pressure, and will be licensed under the GPL and incorporated into
 the ALSA driver suite.  See 
 http://hyppo.screwdriver.net/show.phtml?id=102919
 for the original announcement.

 The Advanced Linux Sound Architecture (ALSA) project 
 (http://alsa.jcu.cz) is
 a project designed to build an architecture for pro-quality sound and
 MIDI applications, from low-level drivers for sound and MIDI hardware
 to high level libraries and sequencers.  The project is committed to
 releasing all work under the GPL.
 
 The 4D Wave chipset is used in the following products:
 
 Company     Product Name
 =======================================
 Addonics    SoundVision (model SV 750)
 AzTech      PCI 64-Q3D
 Best Union  Miss Melody 4DWave PCI
 CHIC        True Sound 4Dwave
 HIS         4DWave PCI
 Jaton       SonicWave 4D
 Paradise    WaveAudio Interactive (Model AWT4DX)
 Promedia    Opera CyberPCI-64
 Shark       Predator4D-PCI
 Stark       PCI
 Warpspeed   ONSpeed 4DWave PCI

 
 You can read more about ALSA and the call to sound card manufacturers
 at http://alsa.jcu.cz/call.html
 
 Thanks,
 Thomas Hudson
 Cygnus Solutions

From: kentdaniels@webtv.net (kent w daniels)
Date: Fri, 5 Mar 1999 14:37:35 -0500 (EST)
To: editor@lwn.net
Subject: Platform support

GIven that Linux is platform-neutral, it is my great hope that as many
Linux users as possible will consider non-Wintel plaftorms in their
purchases.  I think supporting the opposition (AMD, PowerPC, etc.) can
strengthen Linux's market as it can deter the monopolistic business
practices of Microsoft and Intel.  I myself am a Mac user, but also
think Linux very much represents the future.  Apple is likewise
releasing a Unix-based operating system later this year.  This raises
the possibility that these two such operating systems can coexist in
way that could strengthen each other.  I also believe that Linux just
may be the best thing to happen to the Mac plaform, as now consumers
will begin to see Unix in a new light.

Date: Wed, 10 Mar 1999 14:44:58 -0500
From: John Kodis <kodis@jagunet.com>
To: editor@lwn.net
Subject: Editorial on FUD

[ Dear Editor,

  Following are some comments on the spate of "Who do you sue?"
  articles you've cited recently.  Please feel free to use this in
  your editorial column.  Thanks, and keep up the good work. ]

The Linux FUD meme seems to have shifted recently.  In months past,
authors looking for a weak spot in which to jab Linux would bring the
"there's no support" idea to bear.  This theme seems to be dying --
perhaps due to effective refutation, perhaps just due to lack of
novelty.  However, recent weeks have seen a new type of misinformation
rising to the surface.  The current FUD-of-the-week seems to be the
"who do you sue" theme, as typified by the articles you cited from
_CIO Magazine_ and _Internet Week_ (Mar 9).

This concept boggles my mind.  Has there ever been a CIO who
approached his board of directors with an IT business plan along the
lines of:

    "To improve our productivity, we'll standardize on software that's
     known to be unreliable; To insure our long-term viability, we'll
     stick to software that's only available as a binary, so that
     we're completely at the mercy of our software vendors.  

    "This way, if anything goes wrong, we can simply cease operations
     and sit back while our team of lawyers find a way around the
     "We're not responsible" clauses in all our shrink-wrapped
     software, and proceed to sue several of the largest, richest, and
     most litigious corporations in the world for damages."

I wish that authors would learn that writing FUD is like writing any
other form of fiction -- the premise has to be plausible for the story
to be entertaining.

-- John Kodis.

Date: Thu, 04 Mar 1999 22:43:01 -0300
From: Leandro =?iso-8859-1?Q?Guimar=E3es?= Faria Corcete Dutra 
To: editor@lwn.net
Subject: XFree and free riders.

	I wonder if it is Richard M Stallman's (RMS) strong ideas --
and even perhaps sometimes unreasonableness, but I suspect he's right
more often than wrong even about technical issues like EMACS vs XEmacs
and gcc vs egcs -- that generates such hate against the GNU General
Public License (GPL).

	When I read "Liz's XFree86 4.0 session notes" at 
http://lwn.net./1999/0304/a/xfree.html,
the following glared thru the text:

>         The development of X is running into finance difficulties.
> There are very big companies interested in keeping up X, but they are
> annoyed with the "free riders", that so many companies benefit and yet
> don't share in the costs.  I am optimistic that it will continue.  If
> the X effort fails, we'll continue and take over.


	The sad thing is that the free rider problem has largely been
solved by the GNU GPL, and some people refuse to use it because it's
too free or because it's too restrictive.  Even if XFree has
pragmatical reasons for not using GPL, as they need to collaborate
with some stubborn video equipment vendors, the solution has been
offered of GPL'ing everything -- and if the proposal was originally
Stallman's, The Open Group (TOG) had accepted it in principle.  But
the XFree refusal prevented this solution.

	GPL'ing the X Window System would sense because then Sun, IBM,
HP, SCO, Compaq and any other vendors, including the free riders,
would have to face a choice of releasing their source code, falling
behind the latest and greatest releases, or entering into commercial
licensing agreements with TOG, what would very probably supply the
funds necessary for continuing development.

	But it seems that people would rather risk ruining their
life's efforts than saying "I was wrong, perhaps that long-haired fat
RMS guy was right!"  If such stubbornness persists, there are many
GPL'd projects like Y or Berlin which would gladly step into XFree's
place, including backwards compatibility!

	The same solution would apply to TrollTech and its Qt product,
which could be superseded by Gtk.


-- 
Leandro Guimar„es Faria Corcete Dutra
Brasil

Date: Sat, 06 Mar 1999 16:47:19 -0800
To: editor@lwn.net
From: Matt & Kim <patawi@polarcom.com>
Subject: Linux doesn't need Corporations, Corporations need Linux.


Now that all the big companies are (apparently) jumping on the
bandwagon it seems like a good time to say something I've been
thinking about for awhile:


Linux doesn't need Corporations, Corporations need Linux.


Think about it. In less than a decade, Linux grew from an academic
curiosity to a world class operating system with some 7 (8,9,10?)
million users. This happened without major corporate involvement. Sure
RedHat has had a significant impact and many good things are coming
out of their involvement RHLabs. But RedHat is a small corporation
relative to the likes of IBM and Compaq.


Why is there such intense capital 'C' Corporate interest in Linux now?


Because they _need_ Linux, or something like it.  Computer software is
now so large and so complex it is impossible for any single company to
develop, and debug, and market, and support. For that matter, it's
probably impossible for a group of companies to do. They are
effectively at war with each other and (long term) cooperation is
against their nature.

Because _they_ need Linux. Billy Borg is assimulating everything in
sight. At the moment, Linux is the only way out. Linux is 'safe'. It
is not owned by any competitors. It is cheap, both in aquisition and
in total cost of ownerhsip. It has a well established and
communicative network of developers, debuggers, and supporters (as in
assistance) with an impressive track record. It is global and
multilingual -- for both humans and computers.


But doesn't Linux 'need' corporations: to become more than hacker's
tool? to get on the desktop? to become mainstream? to achieve world
domination?


Short answer: No.

Linux is already more than a hacker's tool. I'm not a hacker, I don't
program, I don't run a webserver, or do many other things a
"traditional" Linux user might.

Linux is already making inroads on the desktop. Look at Gnome and KDE.
Sure we're not there yet, but I don't have any doubts we won't make
it, and not too far in the future either.

Mainstream? World Domination? From where I sit, it's _inevitable_.
Linux is, or is fast becoming, technically superior to the
alternatives. The Linux development model is the only one which _can_
effectively tie the world's computer system together.

Corporate involvement could potentially make all these things happen
more quickly. I think it more likely they will slow it down as they
apply different strategies to attempt to own or dominate the biggest
piece of the Linux pie. Just look at the browser wars, MSN, and Java
for prime examples. Coming soon to a theater near you.

If Linux really wants a big boost, it should go after Governments
first.

Nothing in this document is orignal. It has all already been said at
one time or another by various people. I think (obviously) it bears
repeating.

A note on 'Linux'. Throughout this document when I use 'Linux' I
really mean the whole community of people and organizations who are
working on and using GNU Utilities, FreeBSD, XFree86, OpenSource and
all the other libertarian development efforts. Linux currently has the
highest profile and has become, to my mind, the flagship.

That's all,

<a href="mailto:matt.wilkie@gov.yk.ca">-matt</a>

Date: 6 Mar 1999 02:02:12 -0000
From: Eric Smith <eric@brouhaha.com>
To: editor@lwn.net
Subject: Cobalt web server security hole

In your 4-MAR-1999 issue, you quoted a Wired News article:

"Vivek Mehra, vice president of product development at Cobalt, said
 the hole, which could give a hacker access to a history file
 documenting a user's activities, wasn't specific to their appliance,
 but to the Linux operating system."

This statement is very inaccurate, and this is not your error or
Wired's, but rather shows that Mr. Mehra is not well versed in how his
product compares to other real-world Linux systems.  In point of fact,
this weakness *IS* specific to their appliance.

Most Linux-based web servers use the Apache web server.  By default,
they do not serve the user's home directory, so this weakness will not
be available for exploitation.  A sysadmin would have to explicity
change the "UserDir" directive in the Apache configuration file to
make this happen.  If Cobalt's product serves up user home directories
by default, that is a brain-damaged decision on their part, not a
weakness of Linux or Apache.

And this weakness of the Cobalt's default configuration doesn't only
compromise the "root" user; it could be used to attack any user's
account.

The default configuration of Apache (not that of the Cobalt product)
is generally quite good, however, there are optional modules that can
improve on it.  For example, on my server I run a module called
"disallow_id", which I have used to prevent Apache from *ever* serving
files owned by root.

Eric Smith
http://www.brouhaha.com/~eric/

Date: Tue, 09 Mar 1999 11:31:51 -0600
From: Craig Goodrich <craig@airnet.net>
To: Ellis Booker <ebooker@cmp.com>
Subject: Vendor Interests are Driving Linux...

[ref http://www.internetwk.com/columns/pers030899.htm ]

... is an excellent column; it should be required
reading for everyone in this crazy biz.  Thanks.

I do have one minor quibble, though:

> But will Linux find its way into the enterprise and (gasp) the
> corporate desktop? Don't bet on it. Commercial firms are
> risk-averse by nature. They are more than willing to pay an OS
> license for the right to sue somebody if things go wrong.

Well, OK, that's a good restatement of the conventional
wisdom, and it's hard to argue with it.  I've been hearing
it for over a year now, and Lord knows our society isn't 
getting any less litigious.  

On the other hand, though, Microsoft has [whatever -- 85%?
90%? 112%?] of the desktop OS market, with products that are 
by near-universal agreement buggy and crash-prone.  Simply 
running Word 6 for an hour on a moderately complex document
would lock up the machine and mulch the doc file back in
1993, and the newer versions are not a whole lot better,
modulo the more complex documents they're supposed to handle.

Windows NT4, Microsoft's paradigm for the enterprise, 
literally leaves a Navy ship dead in the water.  

Microsoft announces the availability of a fix for a bug that's
been in Win95 since it was Chicago.  The bug crashes the OS
when the 32-bit millisecond uptime counter rolls over at
49.7 days; nobody had noticed it in five years because nobody
_even at Microsoft_ had ever managed to keep Win95 up that long 
in the first place.

And yet -- when was the last time you heard of some corporation
suing Microsoft for lost time and productivity?  Maybe it's
happened, but in over a decade I've never heard of it.  The
corporations have simply learned to accept bugs and crashes
and lost work the way they've learned to accept theft of 
paper clips and legal pads.  And Microsoft's reputation for 
support is hardly anything to brag about....

Now, the PC revolution of the '80s involved a corporate culture
shift in the way computing was perceived.  I can still remember
columns pointing out that when all was said and done, corporations
would never entrust their mission-critical data to desktop toys.  
(Many of them did -- possibly most of them -- and it may have been
a mistake, but that's beside the point.)  Embracing Linux at the
server level, then sneaking it into engineering departments,
may lead to yet another culture shift.  

There are some fairly serious but not insuperable problems 
remaining for Linux on the average corporate desktop --
installation is infinitely easier now than it was when I started
using Linux a few short years ago, but it's still 'way over the
head of users who have never installed _anything_; office 
automation tools are few -- though Applix, Star Office, and 
Corel's forthcoming Word Perfect Suite (not to mention whatever 
Lotus does) will make the effective range of choice wider for 
Linux than it is for Windows, where MS Office so dominates the 
market that it might as well be the _only_ choice.  And then of 
course there's the longstanding problem of X configuration and 
(still!) printing.

But all of this stuff is being worked on, and at a frenetic pace.
KDE is already a polished product; Gnome is coming along quite
nicely (in fits and starts, like everything else in the open 
source field); the various distributions are competing with each 
other in painlessness and elegance of installation, and the Big 
Boys are starting to throw their weight (and their software 
engineers) around.

So I dunno.  Looking back over twenty years as a techie, the 
only dependable truths I'm left with are a) it'll always be
cheaper and faster next year, unless it's from Microsoft, and
b) all predictions about what'll happen next in this silly 
business are very likely to be wrong.

Thanks again for the column, keep up the good work --

Craig

=============================
Craig Goodrich 
Rural Village Systems
somewhere in the woods near Huntsville, Alabama

Politics for the Thinking Redneck -- http://airnet.net/craig/g4c
Linux miscellany                  -- http://airnet.net/craig/linux