See also: last week's Back page page.

Linux links of the week

Finally, this proposal works for the Open Source movement as well, which suffers from the fact that commercial software pays much better. Some see the OSS as a communist movement to deny creative people of the value of their labor and destroy the market for software. PURLS can be used as a solution to the joint tenancy problem for software. Imagine the next project after Linux, in which talented and hardworking volunteers receive (initially worthless) securities for their efforts...

This week in history

Stable kernel 2.0.34 was released, as was development kernel 2.1.101. Cobalt Microserver (now Cobalt Networks) announced that it supports the open source development model. Ralph Nader sent Dell a note suggesting that some people want to buy computers without Windows installed. Gimp 1.0 was released - finally, as was GNOME 0.20.

ZDNet sounded off this week with this column by an author who admits at the beginning that he has never tried Linux:

Yes, the Linux camp, like a fat, speedy penguin, is making noise. However, some of the 25 letters I've gotten on Linux simply refer to it as a Windows alternative, not something they worship or even use. Still, it's easy to like the idea of it.... Linux has a snowball's chance in hell of making perceptible inroads against Windows.

One year ago this week saw the return of an old nasty proposed law called UCC2b under the new name "UCITA." UCITA continues to make the rounds at present, and it still threatens the rights of software users throughout the U.S.

The development kernel was 2.3.4; the stable kernel release remained at 2.2.9. Conectiva Linux announced its Spanish-language version. Linux-Mandrake 6.0 was released. Rasterman resigned from Red Hat after disagreements with his supervisor. ActiveState contracts with Microsoft to port Perl.

And News.com worried about the Linux IPOs which were just rumors at that point:

If a company such as VA or Red Hat went public and made a lot of money off Linux, "What does that mean for all those people who've done a lot of work and don't necessarily" make money out of it? Will they still want to contribute to Linux? "That's one of the issues we're struggling with," [VA Linux CEO Larry] Augustin said.

Letters to the editor

Date: Thu, 1 Jun 2000 18:12:37 -0400
From: "Bill Rugolsky Jr." <rugolsky@ead.dsa.com>
To: letters@lwn.net
Subject: Free Software and Embedded Systems

I strongly agree that freedom is an issue with embedded systems,
particularly as those systems become "ubiquitous," and we have
little choice but to use them.  There are hazards ahead, though.

Certification requirements present a potential roadblock to the
distribution of free software with embedded systems.  This has
already become an issue for the ISDN drivers for Linux -- many
countries place restrictions on what can be connected to the public
telephone system.

In the past, when most device functionality was implemented in
hardware, it was simpler for implementations to be certified as legal,
safe, and interoperable.  Today we have software modems; soon we will
have software-controlled radio (see http://www.sdrforum.org/mmits.html).

How should the FCC classify a handheld communications device
that is capable of morphing into a transmitter or receiver across
a significant portion of the spectrum, or switching between various
signalling standards?  How will it enforce restrictions, such as the
the frequency range restrictions that were implemented in consumer-grade
frequency scanners to prevent eavesdropping on (analog) phone calls?

The issues of security, privacy, and reliability that the digital
revolution has forced us to address will soon touch the remote corners
of engineering.  Ironically, the adoption of software monocultures built
on standards (whether it is TCP/IP, Java, etc.) may increase
systemic risks in the short-run.  With freedom must come a heightened
sense of individual responsibility, as the digital world offers an
individual unprecedented leverage for achieving good or ill.

If we, the engineers and the coders, do not address these issues
adequately, the time will quickly arrive when the politicians,
lawyers, and bureaucrats address it for us.  I doubt we will be
happy with the outcome.

Regards,

   Bill Rugolsky

Date: Thu, 01 Jun 2000 17:39:23 -0500
From: Bret Indrelee <breti@ancor.com>
To: lwn@lwn.net
Subject: Embedded source code

In your editorial about access to the source code for embedded systems,
you failed to mention some of the problems with providing open access to
the code.

Some devices have design requirements and operating ranges that they
depend on software to enforce. If it is open source, what are the legal
issues? If someone changes the embedded code in a device and it causes a
fault, whom is at fault?

Do you really want someone hacking their car's engine control or ABS
system? These are embedded systems.

An embedded system isn't supposed to be a computer. It is supposed to be
a widget that performs a specialized task. The manufacturer of a DVD
player doesn't provide a complete schematic to the buyer, why should
they be expected to provide source code?

-Bret

--
Bret Indrelee        Ancor Communications, Inc.
breti@ancor.com      6321 Bury Driver, St 13, Eden Prairie, MN 55346

Date: Mon, 05 Jun 2000 02:45:32 +0000
From: Oliver White <ojw@iinet.net.au>
To: lwn@lwn.net
Subject: Possibilities for free CASE tools

Whilst the likes of Linus and other veterans of the linux coding camp
have done an impressive job with the tools at hand - gcc, gdb, vi, emacs
and so forth, I really do enjoy using CASE tools to develop software.
That is not to say that I don't enjoy sitting down with vi and hacking
up some raw code, but tools like UML (the object oriented modeling tool)
are very useful in visualising code in early stages. Some people are
fond of other tools, but UML is first on my wishlist.

The most popular UML modeling tool around is Rational's Rose. Whilst
certainly full of useful features, it is far from easily customisable,
and with a price tag of around $2500 it's not something a team of
hackers can use to share and distribute easily modifiable models.

There are a couple of Free alternatives for developing UML models. The
most popular would be the gtk based 'Dia'. Whilst great at diagrams, it
is lacking in a number of ways. The first and foremost is that it is a
diagramming tool, not an object modeling tool. The development team
don't seem focused on pushing development this way. Perhaps they
shouldn't, and continue making it a great diagramming program. This is
all well and good.

The only other product I've seen is ArgoUML. This Java based product is
actually quite good, being able to produce code from diagrams and vice
versa. However, as a Java tool it has many disadvantages. Rather than
list flaws in the platform itself, I'll simply make the point that one
has to install non-free software on one's computer in order to use the
program. This defies the point of having a free product in the first
place, in my opinion.

More to the point, linux developers are far more interested in making
their software interoperable with systems written in C++ than in Java.
Interoperability is one of the main requirements for any UML tool I'm
willing to use. This is the principle of linux development that really
appeals to me - I can tie in any number of components to make my life
easier as a programmer. Indeed, should the dia team create a module
simply for rendering UML diagrams to the screen, this would be quite a
reasonable component for any modeling tool, perhaps using components
from my tool, perhaps not.

What I'm saying here is that as we mature into a massive, distributed
development team, tools like UML and other CASE products we can conceive
of will be necessary for some kinds of systems development. No one would
argue against the usefulness of combining various tools to develop
software, and with free software interoperability is a cherished goal,
rather than something an industry has to have forced upon it by
customers unwilling to standardise on one vendor.

Big companies must spend millions of dollars on CASE tools. If they
would pool their resources they could begin development of free software
CASE tools for a fraction of that cost. It seems like a good investment
to me, anyway. I'd like to hear from companies and distributed internet
development teams on how we could get the ball rolling. Once it starts
it will beget it's own future.

Yours truely,

Oliver White

Date: Thu, 08 Jun 2000 05:33:26 +0530
From: Anand Srivastava <anand@aplion.stpn.soft.net>
To: evan@starnix.com, letters@lwn.net
Subject: RE: Fatal Flaw in BSD

Hi,

BSD License has a big loophole and we have witnessed the consequences for
as long as we can see. If BSD Unix was licensed under something like GPL,
it would have developed like it has always developed, even more so, because
the major modifications would not have gone proprietory. Sun would be a
hardware company. IBM would still be a service company. HP wouldn't have to
go into Software at all. The academic community would have had a voice when
we had those horrible GUI wars.  Linux would not have been required.  Of
course companies may not have used BSD at all instead they may have
preffered SVR4. But some companies (IBM ??) would have discovered the
benefits of software that is developed in an open environment. And the
conditions that we have today would have been then. Maybe Microsoft
wouldn't have developed Windows at all, and would have just done what it
does best ie making Software to run on whatever GUI unix had.

OK this is all surmises, but I expect the software world to have been a
much better place if only BSD was released under GPL. Just imagine that you
had a unix just when those microprocessor kits where coming out.

-anand