See also: last week's Back page page.

Linux links of the week

See also Les Nouvelles Neuves de Linux, run by Stéfane Fermigier and others in France. The site itself is in French, surprisingly enough, but the right-hand side bar contains interesting news links that point mostly to English-language sources.

Section Editor: Jon Corbet

Letters to the editor

Date: Wed, 05 May 1999 17:46:13 -0400
From: Joe Drew <hoserhead@bigfoot.com>
To: drwho@xnet.com, editor@lwn.net
Subject: Re: "Restrictively Unrestrictive: The GPL License in Software"

I disagree strongly with your opinions, but they are your opinions and
you're entitled to them. However, I have some nitpicks:

1) Your political views come through very strongly when you explain
various parts of the GPL. If you were going to revise this document,
one fo the first things you should do is make the entire document,
excepting "My opinions" completely non-partisan - notably words such
as 'infected', etc.

2) Communism and capitalism are not mutually exclusive, and neither is
'right' or 'proper' any more than their suitability for a given
situation.

3) Nowhere does the Free Software Foundation gain any rights to your
code that others do not receive - your assertation that 'you are not
the real owner of your code, the Free Software Foundation is' is
completely false. You are the holder of copyright, and as such you can
sublicense your code under any license you want. Just like most other
licenses, though, you can't revoke rights already granted to other
uses (ie, the rights given under the GPL.)

4) RMS does not wish GNU/Linux to be called GNU/Linux because its
contents are, by and large, GPL'd - he wishes it to be called
GNU/Linux because distributions of Linux - particularly Debian - are,
for the most part, the finished product of the GNU project, using the
Linux kernel.

Now, my opinion:

- The GPL protects your code from becoming proprietary. If you don't
care about that, you wouldn't be using the GPL.

- Nothing is inherently Communistic about the GPL. You're not required
to give out your code or changes, but you ARE required to license any
of those changes under the GPL if you do distribute them.

- The GPL has, and continues to, protect Free Software; it has never,
and will never be, concerned with political extremism. It's the reason
most new Free Software exists.

I hope that you take the time to try to understand why the GPL is so
popular, and also that you will revise your document to remove your
personal opinion from the section meant to simply explain the
differences between the BSD license and the GPL.  --

Joe Drew
http://www.woot.net

Reality must take precedence over public relations, for Mother 
Nature cannot be fooled.
      -- R.P. Feynman

Date: Wed, 5 May 1999 11:57:05 +0100
From: Aaron.Trevena@msasglobal.com
Subject: benchmark - not flame
To: bruce@mindcraft.com, editor@lwn.net 

     Bruce,
     
     I would like to add some comments to your open benchmark project.
     
     - Apache server is not designed to be fast - its designed to be
     versatile and reliable. There is a wide choice of web servers
     designed for speed, flexibility, etc for Unix, Linux and
     *BSD. You should include at least one alternative known for its
     speed.
     
     - NT is used as a client system in enterprise environments
     because, frankly, win 9x doesn't do the job. Therefore an pure 9x
     base is unrealistic because you wouldn't have a quad xeon server
     just for secretaries, who are the only staff with win9x on their
     desktop.
     
     - your hardware is known to be designed to work with NT - it is
     advertised Dell policy, this weighs the benchmark heavily in NT's
     favour regardless of any tuning.
     
     - the hardware being tested is unlikely to be used in a Linux
     environment because it is uneconomical - several mid-range
     servers clustered or loadshared would be more appropriate
     providing better performance and increased reliability,
     scalability and accessability.  just because NT can't cluster (2
     isn't a cluster, its a joke) doesn't mean that other system can't
     make better use of hardware.
     
     - the web serving environment is unrealistic, it bears no
     resemblance to a real world serving environment - be it internet
     or external. A machine with a fraction of the power used in your
     benchmarks would rapidly saturate even multinational companies
     networks, the only need for such hardware would be if there was
     heavy use of dynamic content or web applications.
     
     I hope you can address these problems, or at least make it clear
     in published results that the benchmark is in an unrealistic and
     contrived environment.
     
     Aaron Trevena, Intra/Internet Developer & Administrator. 

Date: Thu, 29 Apr 1999 11:58:50 +0200
From: Hubert Tonneau <hubert.tonneau@easynet.fr>
To: editor@lwn.net
Subject: Refer to history to get the truth

> Now, of course, there are reasons for this behavior. One could
> say, for example, that these companies are simply trying to
> prevent the publication of something like the Mindcraft report
> that has drawn so much scorn over the last couple of weeks.
> There's probably some truth to that. Much bad behavior comes
> as the result of good intentions. But, in the end, freedom is more
> important.

The sentence "There's probably some truth to that" is completely
false: I have never seen any benchmark of both Oracle and SQL server
in any review during the last years. Some reviews have been very
serious, with tests run in relation with various database publishers.
The true reason is that they have a powerfull and well organised
marketing division and they prefer to rely on it to get the product
sold.
Your sentence is what they like to hear because the doubt about their
true reasons is good for closed company, better than the crude
reality. So please refer to history: facts are there.

Regards,
Hubert Tonneau

Date: Fri, 30 Apr 1999 04:41:08 -0700 (PDT)
From: Jonathan Walther <krooger@debian.org>
To: mhegarty@mit.edu
Subject: freedom for ITS source?


Dear Mr Hegarty,

I have cc'd this message to Alan Bawden and Richard Stallman, who may
have some comments of their own afterward, and I am sure, will correct
me where I am wrong on historical details.

What it is: ITS is an operating system, in the same way that
GNU/Linux, VMS, and Windows NT are.  More specifically, it is an
operating system developed more than 20 years ago that ran on special
hardware specific to MIT which no longer exists and which has not, to
my knowledge, ever been used for commercial purposes.

ITS has no license, and without one, cannot legally be redistributed
in source or other form.  In your role as the person in charge of
source licensing in MIT's Technology Licensing department, we would
like to ask that you continue the fine forward looking tradition of
MIT in the field of computer science by licensing ITS under the GPL.

From available facts, the GPL would be most appropriate.  ITS was
developed in response to some proprietary drivers which came without
source, but didn't do what the AI lab researchers needed.  Having an
operating system with its source code open to all meant many
researchers improved it, enhancing the computing experience.  The
whole source code of the operating system was completely open.  It was
this openness, and the culture that came with it, that inspired the
GPL, the GNU system, and laid the foundation that vaulted Linux to
fame.

It's nice to be able to look at ones roots.  ITS has never had a
license: it never needed one.  Its been available to whoever knew
someone with the source code who was willing to give it to you.  It
would be a shame to make it any less open than its successors.

If for some legal reasons we are not aware of, it is necessary to
release the code under a more restricted license, I will still be
interested in corresponding.

This jargon file entry contains some detail on ITS:
http://www.tuxedo.org/~esr/jargon/frames/ITS.html

Yours truly,

Jonathan Walther