See also: last week's Letters page.

Letters to the editor

Date: Thu, 05 Apr 2001 12:59:20 +0000
From: Oliver White <ojw@unite.com.au>
To: letters@lwn.net
Subject: Auto-upgrades and trust

Some very valid points have been made, this week, regarding automatic
upgrades of software. I love debian's upgrade features and frankly, I
consider the risk of getting unwanted software worth it. At the end of
the day, if I have to reinstall my system I'll survive. I don't run
critical services on my box.

The folks at Eazel are looking at creating a business model around
automatic updates (or so you report). Trust is of the utmost importance
to this business model. If they can keep a clean record of user
satisfaction then they'll have a steady income. We can stop being system
administrators (unless we want to be) and get on with the business of
using our tools. This is all very good.

A server administrator might want to use authentication and
authorisation for their upgrades. Someone like myself might just want to
be offered the release notes before allowing the system to upgrade
itself. 

Upgrade service vendors are going to have to keep their customers very
happy if they want to stay in business. I think this fact alone ought to
comfort us.

--
Oliver White
WorldForge Developer
http://www.worldforge.org

Date: Sat, 7 Apr 2001 11:35:42 -0600
From: yodaiken@fsmlabs.com
To: letters@lwn.net
Subject: Kernel hackers meeting


Couple of things that struck me:

1. NUMA discussion. This was not really focused on NUMA (in case anyone cares, I think NUMA
   is good architecture, and cc-numa is a major advance on SMP for _small_ numbers of
   processors -- less than 8). The issue was how to design kernels for big MP machines and there
   is a major argument between people who think that IRIS was a success and should be
   emulated and those of us who think IRIS shows exactly how dangerous the "fine
   grained locking everything" approach to kernel development is.  My position is that
   people should look at other designs like VMS Galaxy and Puma   and at
   the real needs of applications before blindly following IRIS into disaster. 
   I also think that it's very wrong to believe that one can easily add complexity to the kernel
   and make it an option that "compiles out".
   
2. Rik's memory management talk. Critical point: scaling up should be
   also considered in terms of effects on scaling down. The design tradeoffs between
   low-end and embedded and the upper end servers are becoming critical issues and the
   issues raised in the NUMA boff are important everywhere. 

3.  Bitkeeper. Both RTLinux and PowerPC Linux trees are bitkeeper based and have been for 
    some time. 



-- 
---------------------------------------------------------
Victor Yodaiken 
Finite State Machine Labs: The RTLinux Company.
 www.fsmlabs.com  www.rtlinux.com

Date: 6 Apr 2001 22:04:34 -0000
From: Eric Smith <eric@brouhaha.com>
To: letters@lwn.net
Subject: Wind River vs. GPL

Gentlemen,

In the lead story of your 5-APR-2001 issue, you wrote:

> Part of the explanation, certainly, is Wind River's distrust of the GPL.

Not surprising.  In 1995-1996 I worked for a company using VxWorks in an
embedded product.  VxWorks is, of course, not GPL'd.  But the
development tool chain was GCC, GDB, etc.  Wind River had a somewhat
customized version of GDB, which, for instance, was task-aware.  When we
requested the source code to their GDB in order to use it on another
platform, they provided us with only the unmodified, official GDB source
distribution.

When we pointed out that the GPL required them to provide us with the
sources used to build the binaries that they had shipped us, they
refused, saying that their modified version was proprietary.

I left that company before I could pursue the matter any further.  But
others have told me that they've had the same experiences with Wind
River since then.

Apparently Wind River was unafraid of section 4 of the GPL, which
provides that if they copied or distributed GDB without complying with
the GPL terms, their license to use GDB would be terminated, and they
would henceforth be unable to use or distribute GDB in any form.

Sincerely,
Eric Smith

Date: Fri, 6 Apr 2001 17:18:54 -0400 (EDT)
From: Joe Klemmer <klemmerj@webtrek.com>
To: <letters@lwn.net>
Subject: The Stability of 2.4.x


	Regarding the stability of the 2.4.x kernel -

> The 2.4.0 kernel was about as stable as it could have been, really.
> The last set of problems takes a wider community of users to find;
> that's what "dot-zero" releases are for. Every stable kernel series
> has taken a few releases to truly stabilize, and 2.4 is no exception.
> Some rough edges remain, but it's getting there.

	Everyone knows that the "stable" kernel isn't really stable till
it reaches the .10 level.  When 2.4.0 was released I remember a bunch of
us Linux people were all saying the same thing when asked by lay-people if
they should upgrade, "Wait till .10".  Maybe that should be a motto...

	"It's good to go at one dot oh!"



Joe "No, I _don't_ have anything better to do" Klemmer

---
If I actually _could_ spell I'd have spelled it right in the first place.

Date: Tue, 10 Apr 2001 03:06:19 -0700
From: Seth Johnson <seth.johnson@realmeasures.dyndns.org>
To: letters@lwn.net
Subject: Information Producers Initiative


I am starting an project called the Information Producers Initiative.

I have pasted a draft of a basic position paper below.  I would like
very much if I could obtain comments on it, perhaps through this list
and/or other public fora.  It is a very general foundation on which
specific policy positions are meant to be based.

I am presently considering developing commentaries on HIPAA (a federal
law addressing medical records privacy) and the Tasini and Napster
cases.

I am specifically interested in obtaining any information regarding
other initiatives that might be similar to this, and what's been tried
and what happened to these initiatives.

I have set up a list for people who are interested in these matters. 
Subscription is by sending an email saying "subscribe C-FIT_Community"
to ListServ@realmeasures.dyndns.org.

Forward this message freely as you wish.

The text below is also available at: 
http://RealMeasures.dyndns.org/C-FIT

Thanks for your help,

Seth Johnson
Committee for Independent Technology
seth.johnson@realmeasures.dyndns.org


The Information Producers Initiative

A Project of the Committee for Independent Technology


The Committee for Independent Technology holds that a proper
consideration of information-related public policy must focus on what
the state of technology means for all citizens.

We believe that a well-founded understanding of the condition in which
citizens presently find themselves as a result of information
technology, should focus on one fundamental principle.

This principle is that information is used to produce new information. 
To put another cast on the same point, information that is accessible in
whatever form has never merely served the purpose of consumption.  This
may seem to be an obvious point, but when it is considered in light of
the new modes of public access that have developed, and the flexible
means of using information that are now at hand, one sees that this
principle is more important now than it may ever have seemed to be
before.

In the past, only specific groups of people, engaged in specific types
of activities, had their interests assessed in terms of their capacity
as information producers.  The public at large has been treated as mere
consumers of information in many areas, with public policy reflecting
this tendency.

Now, however, we all have the capacity to participate in the development
of human knowledge, on a reasonably equal footing with all other
citizens, because of the forms of access to the public sphere that are
now available, and to the forms of information that may be found there,
by means of public communications networks such as the Internet.  This
puts us all in an entirely new position with respect to our abilities to
access, manipulate and produce information.

We may now manipulate information in a profoundly flexible way.  We may
quickly access any work that is available electronically on public
communications networks.  We may, with great facility, decompose any
digitized work into component parts.  We may manipulate, analyze,
synthesize, select and combine the conclusions, observations, discrete
facts, ideas, images, musical passages, binary bits and other elements
of any information in digital form.  We may efficiently produce useful,
meaningful and creative expressive works on the basis of this flexible
access to information.

But perhaps the most far-reaching way in which information technology
affects our condition as citizens, is in the fact that we may all now
distribute our information products to the public at large in a powerful
and convenient manner that obviates the need to rely on publishers and
other intermediaries who have traditionally provided public access to
information producers.

We must no longer allow our rights in the area of the access to and use
of information and information technology, to be regarded merely as
rights of consumption.  All citizens must assure that policy makers no
longer treat their interests in information merely with respect to their
capacity as consumers.  We must advocate for and guard our broader
interests as information producers in equal standing in the public
sphere, possessing essential powers and rights in the access, use and
communication of information.

The Committee for Independent Technology seeks to assure that the rights
and capabilities of all citizens are not undermined through public
policies that restrict the ordinary exercise of their rights to access
and produce information by flexible means.