[LWN Logo]

 Main page
 Linux in the news
 Back page
All in one big page

See also: last week's Back page page.

Linux links of the week

Open source and open data. This page advocates standardizing the file formats used for many popular applications, such as word processors, spreadsheets, and even databases. Your editor has seen many efforts at standardizing data formats over the years; it is not an easy task. But the rewards can be great if a format is successful. See, for example, the NetCDF format, which has brought sanity to some types of scientific data. It's a prize worth trying for.

The Silicon Valley Linux Users Group gets to have an unfair amount of fun, and their web page reflects it. See, for example, the reports from their recent demonstration at the opening of the Microsoft Silicon Valley campus. "The Microsoft people thanked us afterward and bought a round of beers and sodas for everyone keeping themselves under control."

Those wanting a free software news site with more of a Microsoft-centric orientation may want to check out the Why you shouldn't use Microsoft products site.

November 19, 1998



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.
From: "Will Tsui" <will@netpedia.com>
To: <editor@lwn.net>
Subject: Web Hosting for Open Source Projects
Date: Sat, 14 Nov 1998 23:39:20 -0500


My name is Will Tsui and I am in charge of a site called Netpedia
(http://www.netpedia.com/). We would like to invite anyone developing open
source software to get a complementary subdomain on netpedia.net. This is a
"no catch" deal. We don't require links, advertisements, etc. Here
developers can host their websites and keep the web community informed about
their contributions. If you are interested in this offer to all open source
developers, e-mail me for further info.

	Thanks for your time,

From: "Ricardo Galli" <gallir@atlas-iap.es>
To: <lwn@lwn.net>
Subject: Linux and databases
Date: Thu, 12 Nov 1998 16:04:32 +0100

Dear LWN editor,

    I send you this letter because I see a lot of press articles about =
Oracle and Informix coming to Linux community, but nobody mentions that =
there are other companies supporting and providing very good comercial =
application for Linux OS.

One of these companies is Empress Inc. (www.empress.com) that developed =
a very fast and reliable RDBMS altoghether with a complete development =

We were using it from middle 1996, when I bought our first Empress suite =
for Linux. Since then we are very happy with the product and it =
overpasses in performance other RTDBMS servers.

Our applications span from dynamic web pages (many, many, =
www.atlas-iap.es/kuhn/ just as one example), digital newspapers =
(www.diaridebalears.com) to an AAA server (programmed with embedded SQL =
in C programs) where our Radius and Tacacs+ servers act as clients (we =
provide dial-up Internet access services) of the AAA server.=20

Should be note that both, Linux boxes and Empress servers run very =
stable with almost no maintenance at all. Our first Linux/Empress server =
(an old Pentium 133) is still working and need 0 (yes, ZERO, we just =
have to delete the long apache logs twice a year) maintenance and is =
serving more than 100.000 queries a day.=20

Furthermore, our own AAA server (see some stats at the end), which was =
developed in C using Empress development toolkit never had a problem at =
all, it works on a database with 15 tables, with the bigger one (access =
logs) having more then 2.000.000 records.

Please note that we do not have any commercial relationship with Empress =
Inc. (we pay all of our licences at standard prices). I just thought =
this a good story and that it's very worth to take a look to this =
company that is supporting Linux providing a very good RDMS and a =
complete development toolkit from the very beginning. I feel also that =
they are ignored in Linux press/media.

STATS from the AAA server (running on a Pentium II 300 MHz).

gallir@star:/home/people/gallir > dbstats=20
DBSERVER(862): CPU load (0-1): 0.001434 Sec/ops: 0.0304 PID: 10451
Elapsed: 85 days 22:58:08 Ops: 350789 Childr: 1 Max: 5 Overlds: 0
** Parent Times
User: 00:00:08.69 Sys: 00:03:44.83 Total: 00:03:53.52
** Total Times
User: 01:39:26.67 Sys: 01:18:00.23 Total: 02:57:26.90

Congratulation for LWN, specially for daily updates section. It's the =
best source of up-to-minute Linux information.

Best regards,

--ricardo galli
University of Balearic Islands
Atlas IAP S.L., Internet Service Provider

To: editor@lwn.net
Subject: reply to Kastrup
From: Nathan Myers <ncm@cantrip.org>
Date: Fri, 13 Nov 1998 20:08:51 -0800

[I don't want to start a flame war, but David Kastrup's
long, prominently placed, but badly misinformed letter
demands a response.]

David Kastrup predicts doom ("Microsoft Linux") and denigrates
GNU g++.  While his opinions are easily ignored, the false history 
he has promoted needs to be countered.  He wrote: 

  C++ has hobbled free compiler development: the incredibly 
  complicated language definition has caused gcc development 
  to freeze.  The FSF's non-commercial development infrastructure 
  of the gcc compiler for the comparatively simple C language 
  could not keep up with the complications of the C++ language.  
  This has resulted in the splitoff of egcs, mostly managed by 
  Cygnus, a commercial entity and large-time contributor.

In fact, C++ has nothing to do with the development of the Gcc 
C compiler.  (They share only a back-end code generator.)  The 
slowdown in development of Gcc had nothing to do with C++ -- 
blame administrative and personal problems at the FSF -- but was 
anyway solved in exemplary fashion by the formation of the Egcs 
group.  Egcs is not "mostly managed by Cygnus"; on its steering 
committee of 13 members, I count four Cygnus employees.  While 
the Cygnus employees are supremely competent, they are far from 
alone in their contributions, as may easily be seen by a glance 
at the Egcs web site, even though many large contributions have 
not been acknowledged by name there.

Gcc, as delivered in the Egcs package, is as advanced as any C++ 
compiler available.  Its incredibly rapid development since the 
Egcs group provided a new rallying point stands as a shining 
example of what free software can accomplish.  Make no mistake:
this was not a case of a corporation taking over development from
a failed free software project.  It was a classic case of users
taking the source, and the development process, into their own hands 
when the previous maintainer failed to keep up.  The Egcs group has
more than equal to the "incredibly complicated" C++ language.

Nathan Myers

To: esr@thyrsus.com
From: "Khimenko Victor" <khim@sch57.msk.ru>
Date: Wed, 18 Nov 1998 20:49:14 +0300 (MSK)
Subject: How it's possible ? RMS forgot OSS definition or what ?

"After reviewing the QPL, I find that it meets all the criteria to be
considered Open Source," said Eric Raymond, author of The Cathedral and the
Bazaar paper and joint copyright holder of the Open Source trademark.
"This license should allow Open Source software to compete on features and
polish rather than ideological position, which will be better for the Open
Source community."

How it's possible ???? QPL is NOT compatible with Open Source Definition !

"Caldera Systems is very excited about Troll Tech's decision to publish Qt
under an Open Source license," said Ransom Love, President and CEO of
Caldera Systems. "Customer response and feedback to the KDE desktop
environment included in our OpenLinux 1.3 release has been
overwhelming. Offering Qt under an Open Source license will allow KDE to
stand on its technical merits without causing undue concerns to the
development community over licensing terms."

"SuSE congratulates Troll Tech for their wise decision to put Qt under an
Open Source License. SuSE thinks that this step will (hopefully) unite the
Linux community again," said Burchard Steinbild, Managing Director of
S.u.S.E. GmbH.  "SuSE wishes Troll Tech much success with their products
and hopes their reputation in the Linux community rises after this move,
they have deserved it."

What's this ??? Are you are all blind or what ?
  1) QPL is not OSD-compliant
  2) Even if QPL will be fixed to be OSD-compliant this will not change even
jot for KDE ! To resolve KDE copyright problem QPL should be GPL-compliant,
not OSD-compliant !

The Open Source Definition (version 1.0) http://www.opensource.org/osd.html :
-- cut --
3. Derived Works

       The license must allow modifications and derived works, and must allow
       them to be distributed under the same terms as the license of the
       original software.
-- cut --

-- cut --
3. You may make modifications to the Software. In order to preserve the
   integrity of the unmodified version of the Software, modifications must be
   distributed in the form of patches, and the following restrictions apply to
   each patch:

     a. Application of the patch must not modify copyright notices in the

     b. The patch must be explicitly licensed by the following clauses without
        additional restriction:

          Permission is hereby granted, free of charge, to any person obtaining
          a copy of this patch, to deal in the patch without restriction,
          including without limitation the rights to use, copy, modify, merge,
          publish, distribute, sublicense, and/or sell copies of the patch,
          subject to the following conditions: Any copyright notice and this
          permission notice must be included in all copies or substantial
          portions of the patch.

     c. The patch must include an accurate description of the modification,
        the date of the modification and the author of the modification.
-- cut --

Clearly not OSD-compatible !

For KDE is more important that QPL is not compatible with GPL (and GPL
compatibility is MUCH more restrictive then Open Source Definition; NPL & MPL
are OSD-compatible but not GPL-compatible !!!):
-- cut --

3. You may copy and distribute the Program (or a work based on it, under
Section 2) in object code or executable form under the terms of Sections 1
and 2 above provided that you also do one of the following:

     a) Accompany it with the complete corresponding machine-readable
     source code, which must be distributed under the terms of Sections 1
     and 2 above on a medium customarily used for software interchange; or,

     b) Accompany it with a written offer, valid for at least three years,
     to give any third party, for a charge no more than your cost of
     physically performing source distribution, a complete machine-readable
     copy of the corresponding source code, to be distributed under the
     terms of Sections 1 and 2 above on a medium customarily used for
     software interchange; or,

     c) Accompany it with the information you received as to the offer to
     distribute corresponding source code. (This alternative is allowed
     only for noncommercial distribution and only if you received the
     program in object code or executable form with such an offer, in
     accord with Subsection b above.)

The source code for a work means the preferred form of the work for making
modifications to it. For an executable work, complete source code means all
the source code for all modules it contains, plus any associated interface
definition files, plus the scripts used to control compilation and
installation of the executable. However, as a special exception, the source
code distributed need not include anything that is normally distributed (in
either source or binary form) with the major components (compiler, kernel,
and so on) of the operating system on which the executable runs, unless
that component itself accompanies the executable.

-- cut --

What's this ? You all want KDE problem to be resolved so much that you could
not understood clear English ? Or may be I'm misunderstood something (Enlish
is not my native language after all)...

Date: Thu, 12 Nov 1998 17:04:00 -0500
From: "Zygo Blaxell" <Zygo.Blaxell.zblaxell@nt.com>
To: editor@lwn.net
Subject: Balkanization of Linux HOWTO

[Unlimited unmodified attributed distribution permitted]

Now that Linux is starting to turn the heads of major business entities,
I keep seeing other Linux people asking the question "What should we do
next?" or "What should we do to stop the new Microsoft threat?"  My
answer is simple, direct, and to the point:  "Do exactly the same thing
we've always been doing."  

For those who joined the party late, this is what we've been doing:

1.  Producing the best software.  

2.  Supporting and defining open standards.

3.  Creating replacements for existing closed-source software without
infringing on protected IP rights.

4.  Sharing information as widely as possible, without hesitation, in
hope that it will find the right person to act upon it.



Michael Dell explicitly states that Microsoft pays Dell to pre-install
Internet Explorer and Netscape doesn't, so Dell ships IE by default. 
"Free" software can't compete with that kind of business logic, nor
should it try.  We can expect the last of the large closed-source
vendors to make desparate attempts to gain market share, including such
desparate tactics as paying customers to introduce dependencies on
closed-source products into their critical business systems.  Our
response to that should be to consistently produce better products until
the closed-source vendors run out of new ideas or money.

In the recent past, open standards have had published specifications but
typically only proprietary implementations have been available.  Even
specifications that include sample source code are still not entirely
open, because you can't use the code without a closed-source OS,
libraries, or compiler tools.  Now that we have complete open-source
operating systems, and therefore pure open-source all the way from the
API layer to the bare hardware, it is no longer necessary to have
standards "tainted" by the lack of at least one fully open-source
implementation.  We should advocate that the term "open standard" means
that not only is the specificiation published, but at least one
completely open-source implementation (right down to the bare hardware
for software, or right up to the API layer for hardware) is available.

Given the past explosive growth and continuing persistence of legacy
closed-source software, it follows that many good ideas have previously
been implemented in closed source.  Linux is a re-implementation of
Unix, WINE is a re-implementation of Win32, and so on.  This will be
necessary as long as there exists closed-source software that does not
have an open-source equivalent.  We should educate vendors of
closed-source software about their golden opportunity right now to
re-release their products as open source before competitive open-source
implementations of those products exist.  We should follow that up by
aggressively implementing legally-clean, high-quality replacements for
significant closed-source software from vendors that do not co-operate. 
The message is, was, and always will be "release source or die,"  but
today the message is perhaps a little louder than it was yesterday.

It is good that people with solid marketing, sales, and PR expertise are
entering the Linux community.  Linux doesn't just need software
developers:  it needs technical writers, venture capitalists, lawyers,
advertisers, pure researchers, teachers, guinea pigs, testers,
journalists, managers, and all the "other people" that a big software
company would have access to.  Some of these people are following the
open-source tradition in fields very distant from software:  for
example, last week I read a good introductory document with some
practical tips on doing one-on-one marketing of Linux and another
document that explains how to write a press release.  We still need to
find good ways that people with work to be done can be connected with
skilled people with time to spare, and we cannot possibly be too good at
bringing lots of "newbies" up to speed quickly and painlessly.

We should NOT be:

1.  Trying to control or manage open-source development.

2.  Going out of our way to support legacy closed-source systems.

3.  Trying to rigidly standardize everything, or produce a "one true
Linux" in any form.

Open source software in general and Linux in particular work as well as
they do because every developer has a vested interest--whether
financial, practical or purely social--in the end product.  Enforcing
specific external goals will destroy a productive open-source
development group, or at least limit its capabilities to those of a
closed-source development group.  People should always do their best
work first.  If some particular job needs to be done, then those who
benefit most from that job should put together the resources to have it
done, and those who are good at a particular job should be allowed to do

Widespread support of open standards means that support for
closed-source systems before they are replaced is merely an unnecessary
distraction.  Many closed-source systems do not offer any opportunity
for continued development once the original system has been duplicated,
as the existing installations cannot afford any modifications or
extensions to existing functionality.  In the extreme, closed-source
systems are supported via emulation of the hardware that they run on,
creating a burden for user and developer alike while providing no
benefits to either at all.  The best way to support a legacy
closed-source system is to produce an open-source implementation of that
system or a convenient migration path to a replacement open-source
system.  This liberates the closed-source vendor's customers by breaking
the vendor's monopoly on licensing, support and distribution services,
and allows customers to enjoy reduced costs, better product, and better
support, all at the same time.  Note that Linux has already successfully
replaced closed-source Unix implementations for many purposes and is now
far ahead in some areas.

The computing industry in general--open and closed-source vendors
included--is still decades or even centuries away from achieving a "one
true" anything.  After 40 years we still don't have "one true character
set" (ASCII, ISO-8859-*, Unicode...), so why should we expect a large
API like POSIX, Java, or Win32 to become a permanent standard
overnight?  The best software and hardware available today is still very
limited, and if history is any kind of guide to the future, it will all
be obsolete tomorrow.  Trying to evangelize "one true"ness just imposes
restrictions on people with good ideas, destroys the flexibility that
prevents obsolescence, and pre-empts the benefits of research and
development before they even start.  Even Microsoft, arguably the most
vocal proponent of "one true"ness, is beginning to feel trapped by this
goal as their costs skyrocket while growth declines.  Some Linux users
will need rigid, supportable stability, and some Linux vendors should
provide it to them; however, at the same time nothing should get in the
way of those who want early access to the latest in research and


Eklektix, Inc. Linux powered! Copyright © 1998 Eklektix, Inc., all rights reserved
Linux ® is a registered trademark of Linus Torvalds