See also: last week's Back page page.

Linux links of the week

Notes on libre software is an extensive and growing document on all aspects of free software. It is available in both English and Spanish.

Section Editor: Jon Corbet

Letters to the editor

Date:	Thu, 15 Apr 1999 11:32:41 -0600
To:	rowan@shandwick.ca
From:	Maurice Hilarius <maurice@harddata.com>
Subject: An article:Linux won't have its day yet 

In a recent article at:
http://www.canoe.ca/MoneyColumnsRowan/mar31_rowan.html
You (Mr. Geoffrey Rowan of Shandwick Canada) wrote some comments 
that we felt demanded a response:

Hello sir.
We recently read your article with great interest.

We do, however, wish to respond to some of the comments contained
therein.  Such as:

"In fact, there are very few true anarchists on the Linux
bandwagon. You'd have to classify most of the fervent supporters of
the new operating system as hopeful opportunists. These are people who
see a pyramid scam-like possibility and want to get in quickly, make
some money and get out before everyone else realizes that the Linux
base will never expand broadly enough to support the kind of market it
needs to become something other than a niche technology."

We feel a need to respond to this comment:

We have built and supported Linux based computers for a wide variety
of clients for over 4 years.  These clients include organisations in
Canada such as the University of Alberta, University of Toronto,
Queens University, University of Western Ontario, Telecommunications
Research Labs, the Alberta Government - Dept.  of the Environment, the
Federal Department of Fisheries and Oceans, and so on. In the private
sector we sell to various corporations, including San Francisco stores
( 113 stores Canada-wide) and CN Rail.

 In the U.S. we supply the U.S. Navy, the Department of Commerce, the
NOAA/ATDD weather satellite organisation, The National Institute of
Standards and Technology, Stanford, MIT, and others.

 The list is FAR more extensive than this, but we feel that this small
selection demonstrates the scope and quality of our client base. We
believe that our clients are competent, and not likely to be "taken
in" by some sham. They have bought Linux systems from us for years,
and continue to do so now, in ever-increasing numbers.

  We do not "foist this off" on them, they come to us and ASK for
these products.

Simply put the concept you have stated in your article is untrue,
unfair, and verging on slanderous.

We look forward to your response, and hope that a retraction and
correction may be forthcoming, in at least as public a forum as the
original article.


Best regards,

  Maurice W. Hilarius             NEW!  Telephone: 01-780-456-9771
  Hard Data Ltd.                  NEW!  FAX:       01-780-456-9772
  11060 - 166 Avenue                    email:maurice@harddata.com
  Edmonton, AB, Canada  - T5X 1Y3       http://www.harddata.com

Date: Fri, 16 Apr 1999 12:01:24 -0400
From: Jeff Hecker <Jeff_Hecker@dpc.senate.gov>
To: editor@lwn.net
Subject: Some more comments on NT vs. linux

Greetings,

I just finished reading Hajo Smulders' <hajo@mindspring.com> comments in
LWN.  I've heard similar comments from others.  I'd like to suggest that
these comments indicate a basic misunderstanding as to what an operating
is supposed to do -- at least in my opinion.

Hajo says, "I get a Blue Screen Of Death about twice a week; but that is
usually because of my own stupid programming."  Unless Hajo is writing
kernel modules or device drivers,  what kind of stupid programming
crashes an enterprise-class, high-end operating system?

One of the main functions of an operating system is to separate various
programs from interfering with each other and with the OS.  Hajo's
program might crash,  but it should't take the entire system with it.

I have similar experiences with NT.  Our office runs some software which
is (so far) only available on NT.  On a good week, the system only
crashes once.  The "fault" is with the application which has a runaway
memory allocation error in certain circumstances.  The "responsibility"
is with Microsoft Windows NT to keep that application error from
crashing everything else running on the system.

The all-too-common workaround to this problem is that people don't run
more than one application on an NT box -- and justifiably so.  Microsoft
sales staff even use its unreliability as a sales tool -- "If NT isn't
reliable enough to run your two programs at once,  then buy two!"  And
people do.

I run both NT and Linux servers at two locations.  The server I
mentioned above sometimes runs for days without dying.   Other lightly
loaded NT servers running only file and print service sometimes run for
weeks at a time before failing. The Linux systems, which run everything
under the sun,  run for months without interruption, and then its
usually a power failure.  The current uptime on the Linux machines is
about 260 days.  And the las restart was for a disk installation.

Microsoft's products might be be more colorful and have more sound
effects, and have more lemming sales staff pushing it,  but when it
comes to reliability,  there is no comparison.

Jeff Hecker
Washington, DC

From: Craig Goodrich <craig@ljl.com>
To: sales@varesearch.com
Subject: NT vs Linux on Insanely Fast Hardware
Date: Fri, 16 Apr 1999 10:13:13 -0500

Fellas,

OK, there's been an awful lot of flaming (inevitable
and regrettable on the 'net) and serious discussion
of Mindcraft's Linux/NT benchmarks.  The consensus
seems to be that Mindcraft was reasonably honest 
and fair but that tuning parameters for Linux and its
associated software on such high-end equipment is
hard to come by.

I read about you guys showing off at some Linux
trade fair your cybernetic Godzilla that built the 2.2
kernel in something like 45 seconds, with memory
and FWSCSI that started in Santa Clara and stretched
to somewhere around Albuquerque, and quad Xeons 
whose heat output could handle a small Minneapolis
suburb in February.  So I'd assume that if anyone 
knows all there is to know about tuning enormous 
Linux servers, it's you.

How about getting together with Mindcraft (or some
NT-oriented OEM) and rerunning the benchmark,
so at least both the Linux and NT communities would
have some more realistic numbers on which to base
their advocacy, flames, and (oh, yes, I almost forgot)
actual purchasing decisions?

And, incidentally, how about having some documentation
guy pick the brains of your best engineers to come up 
with a central source for Linux server tuning information?

Thanks,

Craig Goodrich

==============

Freeing software is a good start.  Now how about people?
http://airnet.net/craig/g4c

Date: Tue, 20 Apr 1999 15:11:49 -0300
From: Leandro =?iso-8859-1?Q?Guimar=E3es?= Faria Corcete Dutra 
To: flux@microsoft.com, editor@lwn.net
Subject: Free Software. Is it Worth the Cost? 

	I understand Douglas Boling is writing for Micro$oft, who does
not love free software.  But that gives no right to him to write such
blunders.

	"If intellectual property isn't property, then just what is
property?  Why not just give away cars, houses, and everything else?"

	Intellectual property was never property in the sense of
goods.  It is a concession in form of patent or copyright to explore
something you created for a limited time, so that you would have some
incentive to continue creating.

	The difference between goods property and intellectual
property is that ideas (or software) carry negligible copying costs.
There is no logical reason to create scarcity where naturally there's
none.


	"If they actually succeed in making software free, no one will
be willing to employ them to create a product with no value."

	In fact, most money in the software industry is made in
support.  This won't disappear with free software.


	"If software is free, why does it matter who takes credit for
it?"

	This is for fairness, a greater value in any ethical system
than success.

	Also, it is for the practical purpose of highlighting the role
the GNU Project has in the success of GNU/Linux in particular and free
software in general.


	"I'm not saying that Stallman is anticapitalist, I'm saying
the whole free software movement is."

	That is a lie without fundament.  I'm part of it, and I'm
"capitalism-agnostic".  Capitalism was never a value in itself, it is
just a natural social system that happens to foster (to some degree)
freedom and prosperity.  And nothing in the free software goes
squarely against neither capitalism nor its founding values freedom
and prosperity.


	"I just want the folks who write that software to be
paid—and paid handsomely—for writing it."

	This is obvious.  Even Richard Stallman gets paid for writing
software, as well as many others in the free sofware movement, and
they write only free software.

	If you are still using your logical capacities after so much
richness ambitions brainwashing, you will see that free software will
kill at most 20% of the revenues of the software industry, while
creating great value for everyone else, improving the availability of
technology to poor people and countries, and providing a much better
-- and open -- foundation for the software maintenance and service
industry, as well as for the content industry which depends on
reliable, useful software!


-- 
Leandro Guimarães Faria Corcete Dutra
Brasil
http://www.terravista.pt./Enseada/1989/