[LWN Logo]
[LWN.net]

Sections:
 Main page
 Linux in the news
 Security
 Kernel
 Distributions
 Development
 Commerce
 Announcements
 Back page
All in one big page

See also: last week's Back page page.

Linux links of the week


ClickWalk AS, Oslo, Norway is a company that has worked to develop a better and more visually-oriented interface to websites. Their premiere site provides information on the city of Oslo, Norway (alternatively, the site can be viewed in Norwegian) by taking you on a visual tour, complete with maps. The emphasis is on speed of access, yet the Oslo tour contains over eight thousand photos. And, of course, since we are telling you about it, the site was also developed on a Linux server using Apache, postgresql, and a java servlet under Apache JServ. For more details, check out their kind note in response to our questions. It also contains a full press release with more information on the site.

Section Editor: Jon Corbet

Guest Editor for the Week: Liz Coolbaugh


March 25, 1999

   

 

Letters to the editor


Letters to the editor should be sent to editor@lwn.net. Preference will be given to letters which are short, to the point, and well written. If you want your email address "anti-spammed" in some way please be sure to let us know. We do not have a policy against anonymous letters, but we will be reluctant to include them.
 
   
Date: Thu, 18 Mar 1999 11:43:41 -0800
From: Marko Rauhamaa <marko@calnet28-100.gtecablemodem.com>
To: editor@lwn.net
Subject: Re: Free Software


Sean Hennessy <geishan@ozmail.com.au>:

> I'm an small independent software developer. 
> The company consists of
> me.  That's all just me.  My work provides 
> for my family. I create
> proprietary software.  I am therefore the enemy

> in some peoples eyes.
>
> I like Linux. I like the open source. Hell I'll proberly even
> contribute to Linux.  But to have all software free?? That sucks.

I'm on the dark side of the force as well, and like you, I'm using
free software in my work and in our product. I'm currently the sole
breadwinner for my family.

Moral issues aside, I believe our fates are sealed. In the end, all
software will be free because free software will simply blow away
proprietary software. We are living a transition period during which
numerous profitable, proprietary niches can be found, but it's only a
matter of time before most applications are free.

You and I may have made our fortunes before the transition is over,
but don't count on it. Whole industries disappear and that affects the
livelihoods of thousands or even millions of people. There's little
you can -- and probably nothing you should -- do to slow the natural
evolution down.

In the future, software developers' work will be like scientific
research:

 (1) They will advance in their careers by publishing their works.

 (2) They will be hired by universities, foundations and governments.

 (3) Their will be paid poor salaries, and unless they are tenured,
     their job security will be nonexistent.

 (4) Their results will kick ass.

I used to be a computer scientist. People in the academia are willing
to work below minimum wage, without pay or as department secretaries
just to be able to continue their work in the research community.

It's a bright future for software.


Marko

-- 
Marko Rauhamaa   mailto:marko.rauhamaa@iki.fi   http://www.iki.fi/pacujo/
Suomenkielinen esperantokurssi http://www.iki.fi/pacujo/esperanto/kurssi/
    Free Esperanto Course http://www.iki.fi/pacujo/esperanto/course/
   
Date: Thu, 18 Mar 1999 23:12:18 -0700
From: Jeffery Cann <jcann@techangle.com>
To: editor@lwn.net
Subject: RE: Free software


--------------65B851295062EF233B499D30
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

In his letter to the LWN Editor, geishan
<geishan@ozemail.com.au> writes:

    "I'm an small independent software developer. The company consists
of me.  That's all just me.  My work provides for my family. I create
proprietary software.  I am therefore the enemy in some peoples eyes."

My first point is that few in the Free Software community would
consider you "an enemy".  They would encourage you to release your
source code under the GNU General Public License (GPL).

He continues his letter with:

    "I like Linux. I like the open source. Hell I'll properly even
     contribute to Linux.  But to have all software free?? That
     sucks."

My second point is that the writer misunderstands the meaning of the
word 'free'.  Richard Stallman has spoken of this problem before and
there is simply no other way to explain it in English.

If you were to release your code under the GPL, that does not mean you
could not profit from it.  In fact, you could sell your application.
You could sell services that relate to you application.  You could
contract yourself to write enhancements for your application.  The
point is that many programmers (including ISVs) have released their
code under the GPL and have made significant profits using the ideas
suggested above.

When released under the GPL, your source code would be free.  Not in
cost, but in libery.  Your code would be free for other programmers
like myself to contribute to the development of your software.  My
development contributions could possibly earn you money, but I would
not contribute my time and programming effort solely for you to
profit.  I would be contributing to your source code because it would
benefit me directly in some other way.  For example, fix an annoying
bug or add an enhancement that could save me hours.

If I have no access to your source code and there is a bug or
enhancement that I would like, I would be helpless to resolve the
situation -- i.e., I would literally be at your mercy for resolution.
Thus, your proprietary license restricts my freedom to resolve this
situation.  This is the meaning of the word 'free' in the GPL.

Note that by enhancing the freedom of your users, you are not limiting
your own freedoms.  You would have the freedom to profit from your
efforts.  There are no such restrictions in the GPL.  On the contrary,
if you read the GPL and other material on the Free Software
Foundation's (FSF) web site (http://www.fsf.org), it clearly explains
how a person can profit from using the GPL.  The additional benefit is
that if you release code under the GPL and later decide not to
maintain it, no one will be left without their freedom to continue to
maintain or enhance the code.

As the ideals of the Free Software Foundation are exposed to a greater
audience, I fear that people will continue to misunderstand what is
meant by free software.

I encourage you not to restrict the liberties of your users (their
freedom) by continuing to release restricted software programs.  I
encourage you to read the GPL and its usage 
on the FSF web site (http://www.fsf.org).

Sincerely,
Jeffery Cann

--
"Who does not trust enough will not be trusted."

- Lao Tsu

   
From: guylhem@barberouge.linux.lmm.com
Date: Thu, 18 Mar 1999 17:04:52 +0200 (GMT-2)
To: editor@lwn.net
Subject: "Libre" better than "free"

I have to agree with Conrad idea, "libre" will avoid confusing free
software, shareware, freeware, gratis...

* Free software would be software you get for $0 like internet
explorer

* Libre software would be software you can freely change and
redistribute without restriction, like what BSD, GPL and many other
licences provides

* Open source is yet another idea, anyone can see/change the sources
but there might be additional restrictions, like QT license or MPL

But who will want to make words clearer for newbies ?  Who can
decide how we should call the kind of software we're talking about ?

We some tough opinion.
   
To: editor@lwn.net
Subject: Standish
From: Nathan Myers <ncm@nospam.cantrip.org>
Date: Thu, 18 Mar 1999 17:40:38 -0800

To the editor,

LWN wrote:
> "The Standish Group recently interviewed Fortune 1000 IT executives
> in several industries. Despite all the hoopla about Linux, not one
> company was found to be running a mission-critical application on
> Linux."  Surely there must be somebody out there with an example that
> can put an end to this?

I'm not sure it does any good to participate in things like this,
even so far as to call for counter-examples.

The Standish Group gets its money by mirroring the biases of its
customers, who are now heavily invested in legacy software.  At the
moment, those customers are paying to be reassured that they haven't
wasted their money.  In the future they will pay to be told
otherwise, and then the Standish Group will oblige.  It will be
their well-paid job to gather examples.  Why should LWN readers do
their work for them, unpaid?

Free Software achieved its current success despite the Standish
Groups of the world.  While the buyers of reports like The Standish
Group's are keeping their ears plugged, Free Software is
strengthening its position.  It is better that they remain
complacent while Free Software infiltrates their organizations from
below, so that they find themselves already dependent on it when
they do wake up, and less inclined to attack it.  Organizations
already using Free Software openly have a competitive advantage
against those using buggy legacy systems, and it benefits Free
Software to preserve that advantage.

The only area I know of where this argument doesn't apply is in
government.  Those reports that influence government administration
decisions need the best up-to-date information, and readers of LWN
can really help in getting those reports into line with reality.

Nathan Myers
http://www.cantrip.org/

   
Date: 18 Mar 99 21:12:53 PST
From: Ken Engel <kenengel@netscape.net>
To: editor@lwn.net
Subject: What's wrong with proprietary software?

By no means am I the first nor the last to have a similar experience
to which Richard Stallman had with Xerox. But I'd like to see more
public awareness (read: raising hell) about this kind of problem.

The company by which I am employed has decided to standardize on a
third-party product. It has been my assignment to create a custom
application with this product. It's been a nightmare.

The third party's primary target platform is Windows NT, and has
been ported using one of those NT-to-UNIX conversion tools. The
result is a slow, bloated, black-box monolith. Proprietary software,
binary file formats and lack of command-line/scripting awareness
constrain my effort to Point-&-Click, Drag-&-Drop
"development". This translates to inefficiency, high cost and high
blood pressure.

Software whose source code is freely available is not merely
ideology - it is the ultimate practical solution. I urge IT
supporters, developers and managers, anyone concerned with the
bottom line, to seriously consider it.

Thank you
Ken Engel
My opinions are mine only.

____________________________________________________________________
More than just email--Get your FREE Netscape WebMail account today at htt=
p://home.netscape.com/netcenter/mail
   
To: editor@lwn.net
Subject: Apple and Open Source.
Date: Tue, 23 Mar 1999 21:05:37 -0400
From: Humberto Ortiz Zuazaga <zuazaga@coqui.net>

Let me get this straight:

1 Apple takes Mach and BSD 4.4 and Apache, bundles it with a GUI

2 Apple releases this as MacOS X server, a commercial product

3 Apple then takes the GUI back out, and releases Darwin under a
license that bright people can't agree whether it's free or not.  In
any case contributions to Darwin can get rolled back into Apple's
proprietary products.

This advances free software how?

In fact this is the kind of thing that the GPL was designed to
prevent.

This is the best argument ever that RMS's presence is still needed.

So what about mklinux?  That's Mach plus a linux single server,
right?  Why didn't Apple ship a product based on that (hint: the
kernel work would be covered by the GPL)?

-- 
Humberto Ortiz-Zuazaga - zuazaga@coqui.net 
http://home.coqui.net/zuazaga/


   
Date: Thu, 25 Mar 1999 11:55:35 -0500 (EST)
From: Conrad Sanderson <conrad@hive.me.gu.edu.au>
To: matthew@mattshouse.com
Subject: comments on your "Here Comes Da FUD" article + more


[ originally to matthew@mattshouse.com ]

An interesting article, pointing out some of the avenues that
Microsoft can take.

What surprises me more is the responses that you are getting.  It
seems there are too many naive people out there thinking that Linux
is half-invincible because it is a "movement" or that Microsoft will
hold itself back, or that other companies will come to the rescue of
Linux.  Microsoft is a 800 pound gorilla that is pissed off.  Most
companies would rather get out of its way than defend something as
small as Linux.

Microsoft doesn't play by the rules - your readers think that
certain facts exclude Microsoft from doing something nasty,
particularly the PR aspects.  But the truth of the matter is that
Microsoft is stupid, arrogant, and above all, a bully.  It hasn't
grown up yet, as is demonstrated by the personality of Bill Gates
(especially in the DOJ video), and as such, will use any means
necessary, whether it is bad PR or not, to attempt to crush or
minimize the "Linux threat".  Microsoft is a teenager in the
computing business, who refuses to grow up.  It doesn't think too
far into the future, and it is ruthless.  The only way it will learn
to be nice and think of the long term effects is when it gets a
bloody nose.  The DOJ lawsuit is only the start.  Something else
besides the lawsuit and Linux is needed to smash Microsoft's face to
a pulp.  What that is, I don't know yet.

Microsoft will attack.  I wish I was wrong.


To learn more about the corporate culture of Microsoft, head over to
"Why you shouldn't use Microsoft products", at: 
http://wave.me.gu.edu.au/~csand/md/0soft.html

Conrad Sanderson - Microelectronic Signal Processing Laboratory
Griffith University, Queensland, Australia


   
To: editor@lwn.net
Subject: Linux needs a registry?
From: Alan Shutko <shutkoa@ugsolutions.com>
Date: 18 Mar 1999 10:53:45 -0600

In his letter on March 16th, Tom Atkinson said that Linux needs a
single database which holds all configuration info, and it Linux
gets one, it'll be easy to write unified configuration managers.
In other words, we need the same kind of thing that NT has in its
registry.  This could be helpful in some ways, but it won't cure
all the problems facing admin tool authors!

There are several issues in configuring other apps:

* Finding the config files.  A central database would help with
  this, but finding config files isn't that much trouble now.

* Reading the config files into some internal structure.  A central
  database would make this much easier, and this would help things
  out considerably.  But it won't do anything for the bulk of the
  labor:

* Dealing with the application specific configuration information.

For example, which is more of a problem configuring sendmail: the
fact that it has a wild configuration file format, or the fact that
it has truly hairy capabilities in that configuration file?  Any
tool which attempts to configure sendmail will have to deal with
its multiple tables, rules, etc regardless of the file format.  You
can write a 4-function calculator using sendmail rules... a config
tool to deal with such information is going to be difficult no
matter what.

For evidence that a central database is no panacea, look at the
Windows Registry.  Lots of information is stored in a central
location, but there are a number of registry entries which store
completely incomprehensible strings.  A Linux config database could
solve some of these problems, but there are difficulties for config
manager authors that the database will not address.

-- 
Alan Shutko <shutkoa@ugsolutions.com> - (314) 344-5214
   
From: Art_Cancro@uncnsrd.mt-kisco.ny.us (Art Cancro)
To: editor@lwn.net
Date: Thu Mar 18 13:27:29 1999
Subject: Dear Editor:        

Dear Editor:  
   
   I'm certain that I speak for a huge segment of the Linux
community when I say that I take exception to Tom Atkinson's
proposal that the various configuration files present on a Linux
system be combined into one huge database.
   
   The configuration database that Mr. Atkinson proposes is,
essentially, the Windows Registry.  Anyone who has had even minimal
experience with the administration of Windows systems knows that
it's quite easy for the registry to become corrupted, or (perhaps
even worse) loaded up with defunct settings for programs which are
no longer installed on the system.
 
   
   While there are advantages to such a database -- the ease with
which it can be saved and restored comes to mind -- to see the many
disadvantages, one need only spend some time running an operating
system which already implements it this way.
   
   Perhaps we can meet somewhere in between.  Configuration files
in a standardized location and with a standardized format, perhaps
using a standard set of library functions to read and manipulate
them (much like pre-registry versions of Windows used standard
functions to access .INI files) might be a worthwhile mix of
flexibility/reliability with easily implementable administration
tools.
   
   ---  
   Art Cancro                            UNCENSORED! BBS  
   ajc@uncnsrd.mt-kisco.ny.us            http://uncnsrd.mt-kisco.ny.us  



   
Date: Fri, 19 Mar 1999 01:37:14 +0100
From: Tom Simonsen <tomsim@eunet.no>
To: editor@lwn.net
Subject: Re: Subject: What Linux needs next

Hi,

Having read the letter from Tom Atkinson about configuration data;
I have to disagree. His idea about a central database for all
configuration data is, in my opinion, a flawed one. This is the
same as the registry in Windows.

A central registry has one major flaw (at least in the MS flavour);
you computer dies when it gets corrupt (note when, not if...).

Tom does not miss the boat though, as he points to a real problem.
Configuration data should follow a standard set of rules; lessening
the learnig curve.

My suggestions are: Use ASCII files, so you can get at the data,
even when booting a minimal system.  All applications should have
sensible default values, or die gracefully, if the configuration
data is missing.  All applications/deamons etc. should have their
own configuration file, so that screwing up one file, doesn't screw
up anything else.  User setup should be in own files, separate from
the system setup

It might be useful to have a central database that works as a cache
though... If things stops working, delete the cache, and it
rebuilds automatically over time.

Thanks for your time,
Tom

   
Date: Thu, 18 Mar 1999 17:31:27 -0800
From: Jeffrey.M.Bolden@healthnet.com
Subject: Tom Atkinson
To: editor@lwn.net

I just read Tom Atkinson's piece on system administration.  I found
it rather ironic to read a piece about how Linux could defeat Win
95.98/NT by adopting a "system database", an approach that amounts
to adopting the Window's registry.  As anyone with experience with
the registry can tell you maintaining and modifying the registry is
a nightmare compared to those ASCII files in Linux.  I imagine
Mr. Atkinson doesn't realize this because he has never tried.  I'd
recommend that he open up regedit and see his unified non-ascii
database in action.

The difficulty in installing Linux does not lie here.


   
Date: Fri, 19 Mar 1999 18:35:26 -0500
From: Herschel Cohen <hcohen2@home.com>
To: editor@lwn.net
Subject: What Linux needs next ... [Not so sure]

Response to Tom Atkinson letter in 18 March, 1999 Linux Weekly News

Tom,

While in general I agree with many of your statements, I disagree
with certain specifics that are myths that are too easily taken at
face value.

For example, Win95 seems easy because it comes pre-installed, read
the documentation on reinstalling Win95 on say my Micro Xku.
Compare that to a new install of NT Workstation, the latter
(<I> i.e. <B> NT </B></I> is <B>
easier! </B> Moreover, configurations under Win95 for such
<I> simple things </I> as sound volume are easy to set
in multiple locations [this time using my laptop as the example],
but one just changes the sound level for that session [Fn+F7]
whereas the sliders are more permanent.

The registry is a database, but what a horror - subject to
corruption newer versions overwrite <B> parts </B> of
older applications.  Fragments of old applications remain to cause
conflicts.  Etc., etc.

I just think Win95 is a poor model.  Nonetheless, I am in the
process of solving problems configuring my latest installation of
Linux.  While I am much further along, I too am convinced that the
vision of casual users flocking to Linux is not practical for the
present condition of this OS.  Moreover, even with OEM
installations and configurations adding a new application that
requires su user access rights may go astray!

So perhaps now you see why I understand and even support some of
your ideas, but I think the solutions cannot be postulated on a
false premise such as the <I> usability of Win95. </I>

Nonetheless, an interesting take on the problem of making the Linux
desktop a more mainstream option.

Take care, Herschel
B/ST Software Developers
   
Date: Tue, 23 Mar 1999 05:38:14 +0100 (CET)
From: "Anthony E. Greene" <agreene@pobox.com>
To: editor@lwn.net
Subject: Re: What Linux Needs Next

In last week's Letters to the Editor, Tom Atkinson advocated a
central database of configuration settings for Linux. This is
supposed to help ease end-user configuration of Linux.

I think he's looking at the problem from the wrong angle. The problem is
not that configuration data is stored in text files. I see the problem as
twofold:

	1. Many of the config files use different formats. This
means a user cannot count on a specific syntax every
time. Sometimes a colon separates the setting key from the value,
sometimes an equals sign, and sometimes just whitespace. Sometimes
multiple values are separated by commas, and sometimes by
whitespace. Comments are usually marked with a pound sign, but
Samba uses a semicolon. User's need to have a consistent interface.

	2. Todays casual users are accustomed to point and click
configuration. A consistent text file syntax would be helpful, but
not nearly as much as a point/click interface, which tends to
enforce some consistency just by having a familiar set of
widgets. This could be accomplished without a consistent text file
syntax, but would be much easier if the underlying text files were
consistent too.

The strength in using separate text files is that a mistake in on
application does not cripple the system. What would happen if the
GUI config utility made an error writing a system-wide
configuration database?  Anyone who has had to recover the Win95
registry can confirm that it can be painful. A distributed database
(text files) avoids that single point of failure.

The other strength of text files is that you can use any text
editor to update the system. Relying on a specific configuration
utility is a single point of failure too. I loved being able to
edit INI files in Win3.11. You could often change settings that
were not available in whatever GUI config utility included with
Windows or with the application. This is an advantage, not a
weakness.

Mr. Atkinson is correct in that it takes some defensive programming
to have applications read hand-edited text files. I see this as an
acceptable price to pay for the resulting flexibility and
reliability.

--
Anthony E. Greene <agreene@pobox.com>
Homepage & PGP Key <http://www.pobox.com/~agreene/>


 

 

 
Eklektix, Inc. Linux powered! Copyright © 1999 Eklektix, Inc., all rights reserved
Linux ® is a registered trademark of Linus Torvalds