It's official: the Red Hat Linux product is no more. The changes announced
by the company can be found discussed, in detail, on
the Red Hat Linux Project page. In
summary, the changes that have been announced are:
- The Red Hat Linux product will no longer be available as a box on
store shelves. Not even in virtual stores. The various Red Hat
enterprise products remain, but the low-end distribution as a
commercial product from Red Hat is done.
- Development of Red Hat Linux will continue, but the company is trying
to move the development of the distribution into a more
community-oriented mode. The internal development mailing lists will be opened
up, and there will eventually be a way for external maintainers to
contribute fixes and packages.
- Red Hat Linux will become more volatile. There will be a six-month
release cycle, with no real distinction between major and minor
releases. Red Hat will stop backporting security fixes to the version
of the relevant package shipped with the distribution release;
instead, applying a security fix will mean upgrading to the latest
version of the affected program. Red Hat will also work harder at
pushing fixes back "upstream," rather than carrying patches
themselves.
There are a few implications of this change for Red Hat Linux users.
Essentially, if you use Red Hat Linux, you will have to pay more. Either
you pay more cash by moving up to the enterprise offerings, or you pay more
in effort by finding bugs in the distribution, and, if you can, helping to
fix them. A Red Hat Linux box has traditionally been a great bargain: a
relatively small amount of money for a stable, well-engineered distribution
containing millions of dollars worth of software. Red Hat Linux will
remain a good deal, but the terms of the bargain are changing a bit.
For high-clue users who would like to be a part of the distribution
development process, the changes will certainly be a good thing. Red Hat
has traditionally been developed in a relatively closed mode. Every now
and then a new release would show up, but the process by which the
development came together was distant and opaque. This distance is one of
the reasons why many hackers have preferred more community oriented
distributions, such as Debian, Mandrake, or, more recently, Gentoo. Red
Hat clearly hopes to tap into the development community by opening things
up in this way. If things go well, the result could well be a better, more
quickly evolving distribution.
Other users will have to think about whether they want to download and
manage new releases themselves, buy a boxed copy from some other retailer
(the number of such products is certain to increase), or switch to a
different distribution. All three are good options, including the last
one. One of the great benefits of using Linux is that you can
switch to a different vendor if you don't like where your current vendor is
going.
This change is a big step for Red Hat; the company did, after all, get its
start by selling boxed Linux distributions at retail. As Linux and the
market have evolved, it has become clear that the retail channel is not
where the real money is to be made. Red Hat, being a public company
needing to bring in serious revenue, is focusing on the markets that, it
hopes, will keep it going. So retail sales are out. But Red Hat cannot
afford to lose its base distribution and the many people who help test it.
Thus the Red Hat Linux Project. With luck, Red Hat can have it both ways:
serious revenue from the enterprise market while building a larger
development community.
Comments (19 posted)
[This article was contributed by Joe 'Zonker' Brockmeier]
As expected, SCO trotted out a new licensing program today that would
give Linux users a license from SCO to what SCO claims is their
intellectual property in the Linux kernel from 2.4 on. SCO also
announced that they had received copyrights for the Unix System V source
code.
I sat in on the teleconference that SCO held to announce the new
licensing program. McBride did most of the talking during the call, with
David Boies adding just a few comments and clarifications, and answering
a few questions that were directly addressed to him. I tried to get in
the queue to ask a few questions about the impact of the GPL on their
plans to offer a license relating to the Linux kernel, but I was not
called on. Don Marti, of Linux
Journal did get a question in about whether SCO would offer any
additional evidence to substantiate their claims, but it was mostly
ducked by McBride, though he did affirm that they were not talking about
code coming from BSD.
During the call, McBride claimed that "hundreds" of files related to
SMP, NUMA and read-copy update (RCU) were infringing on SCO IP either
directly or indirectly. According to McBride, if the Linux community
were to remove the offending code there would be "little non-infringing
code" left in the areas that SCO is claiming rights to. Essentially, SCO
seems to be basically claiming ownership of most of the advancements in
scalability whether they are directly taken from SCO's codebase or not.
Also, McBride noted that some of the code that they claim infringes on
their IP was not contributed by IBM, though he did not specify which
vendor(s) he believed to be responsible.
Other than announcing the new plan and the copyright registration, very
little information was forthcoming. Essentially, they intend to offer a
license of some kind that would idemnify companies from possible suits
for copyright violations. Pricing was not disclosed, though McBride
hinted that it would be equitable or similar to UnixWare 7.1.3
licensing. It will also likely be a per-server, per-CPU situation.
Though SCO did not disclose all of the license terms today, it doesn't
seem possible that the company would be able to abide by the terms of
the GPL while charging for licenses to run their IP in conjunction with
Linux. Even if SCO actually legitimately holds claim over code that's
being used in the kernel, the voluntary act of licensing that code
should require SCO to allow distribution of the same code under the GPL.
According to Section 2b of the GNU GPL:
You must cause any work that you distribute or publish, that in whole or
in part contains or is derived from the Program or any part thereof, to
be licensed as a whole at no charge to all third parties under the terms
of this License.
And, if that weren't enough, Section 4 enjoins anyone from sublicensing
programs under the GPL:
You may not copy, modify, sublicense, or distribute the Program except
as expressly provided under this License. Any attempt otherwise to copy,
modify, sublicense or distribute the Program is void, and will
automatically terminate your rights under this License. However, parties
who have received copies, or rights, from you under this License will
not have their licenses terminated so long as such parties remain in
full compliance.
Even "hundreds of files" would still be considered a derivative of the
Linux kernel -- the majority of which is still uncontestedly free and
clear of SCO's IP. If you take the folks from SCO at their word, and
assume that they really do own claim to these "hundreds of files,"
they're ultimately useless without the remainder of the Linux kernel --
which is still under the GPL.
And it's far from clear that SCO has legitimate claim over any code
being used in the Linux kernel. Unfortunately, but not surprisingly, SCO
did not directly address the issue of how they could license code that's
already been distributed as part of the Linux kernel as a separate
component that would not fall under the terms of the GNU General Public
License.
McBride also made a point of emphasizing that the SCO license would be
for "binary format." Which is puzzling, since SCO does not seem to be
offering to distribute any kind of new code or new kernel -- simply a
license that would give SCO's blessing to using code already available
in the Linux kernel. McBride made the point several times that SCO would
not be offering source code licenses.
While SCO's antics have most of the Linux community seeing red,
someone out there is responding well. SCO's share price has jumped more
than a dollar today, and looks likely to close above $13 for the first
time in a year.
On the IBM front, Boies was asked whether there were any new
developments in SCO's case against IBM. Boies said that he had "nothing
to add" and that "as a litigator, I assume cases are going to court
resolution." Boies also noted that a lack of resolution in the IBM case
will not stop SCO from going forward with other plans based on claiming
IP infringement in Linux.
In a nutshell, SCO is formalizing a plan to try to charge companies for
the privilege of using Linux or sue them for not doing so. Whether
companies will be willing to do so remains to be seen. If they do so, it
will basically be on SCO's say-so that they own the rights that they are
trying to sell.
Comments (50 posted)
It has now been one year since we posted, in
the
July 25, 2002 Weekly Edition, the notice that LWN.net was to be shut
down as a result of its financial problems. The staff had been working
without salaries for months, and nothing we had tried seemed to work. It
was a hard thing to face, but the only option we seemed to have was to pull
the plug. As we said:
This has not been an easy decision to make, to say the least. But,
barring some sort of last minute miracle (do contact us if you have
one, please!), we do not see any alternative.
What happened then, of course, can only be described as a last-minute
miracle. LWN readers started making donations at levels we had never seen
before, and we decided to rethink things one more time. What we came up
with is the subscription scheme which has supported LWN over the last
year.
It is hard not to feel good about how far LWN has come. It is paying its
bills (as long as we keep the bills small), and we
are answerable directly to our readers. We have managed to make a number
of improvements to our content and to the site; traffic is at an all-time
high. In some ways, LWN looks more healthy than it has ever been.
Certainly, we are glad that LWN did not shut down after all.
That said, it is important to note that the problem is still not completely
solved. Salaries remain low, and money for things like travel to trade
shows (important in this line of work) remains scarce. We also very much
need to bring in one more editor to fill out the content and make it
possible for the rest of us to take an occasional vacation. That editor
remains a distant dream for now, however.
What we need to do, of course, is bring in more subscriptions. Recent
changes have helped in that regard; the number of subscribers has been
going up after a few months of little change. We are working at actively
promoting the site - for the first time in its history - as a way of
bringing in more readers. We'll get there. Meanwhile, we are glad to
still be here. Many thanks to all of LWN's readers; your support for us
over the last five and a half years has been amazing. Miraculous, even.
Comments (19 posted)
Page editor: Jonathan Corbet
Inside this week's LWN.net Weekly Edition
- Security: Honeytokens; new vulnerabilities in the 2.4 kernel
- Kernel: The 2003 Kernel Summit
- Distributions: A Look at Arch Linux; new Red Hat beta; SuSE 8.2 review
- Development: Samba 3.0.0 Beta 3 available,
New versions of JACK, MySQL, Zope, MusE, Galeon, Mozilla,
XFce4, Samba, Bluefish, Gaim, Python.
- Press: SCO's Linux Licensing Program, Saving the Net, Linux in Afghanistan, Koha
in France.
- Announcements: Desktop Linux Consortium launches, changes at Mozilla,
PerlBugathon Results, Plone Conference, kdedevelopers.org.
- Letters: Regarding SCO: What are we waiting for, What to do about the RIAA.
Next page:
Security>>