[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

The Unix Guru Universe ("UGU") is a comprehensive collection of information relating to the administration of Unix-type systems in general. They have a largish Linux section, but much of what's there is not Linux specific (but useful anyway).

Designing a Linux PC takes you step-by-step through the process of choosing your hardware wisely. Great levels of detail are not present here, but there is some good advice on what to get. (Another good source for this sort of information is Net Express. They are a high-end Linux systems VAR, and they go to great lengths to tell the reader about why they choose the hardware they do. You can never buy a thing from them and still benefit from visiting their site).

October 1, 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.

We got some complaints about the smaller font used to format the letters last week, so we've done away with it. We'll work on the details of the formatting later... Letters this week cover UDI, Acorn, and some last RMS notes.

Date: Tue, 29 Sep 98 21:27:33 EDT
From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
To: lwn@lwn.net
Subject: UDI comment

Thank you for the coverage and commentary on the UDI effort.

I agree with your assesment that this is likely not good for the Linux
community.  A vital element of Linux's success has been its excellent device
support.  I don't see UDI enhancing that.  Rather I see UDI as a way for
other OSes to gain additional device drivers without providing anything in
return.  It will also encourage binary-only device drivers from vendors in
place of open source support.

What do we gain from UDI?  What other Unix has such extensive device
support?  Certainly not Solaris or BSD.  And they are unlikely to convert
their proprietary drivers to an open interface.

Finally, when there is a defined binary interface, the GPL can be
circumvented with the claim that a device driver is just another stand-alone
program that may be ported to another system.  

Donald Becker					  becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center,  Greenbelt, MD.  20771
301-286-0882	     http://cesdis.gsfc.nasa.gov/people/becker/whoiam.html
Date: Thu, 24 Sep 1998 11:41:32 -0700
From: peter.m.eggers@boeing.com
To: editor@lwn.net
Subject: UDI

I think Intel has a good idea.  To realize UDI on Linux, Intel should
volunteer to become the project coordinator and get the other commercial
UDI members to provide programmer support to create the UDI drivers.  
The Linux kernel development team I am sure would appreciate having but
one commercial entity to deal with and I believe that Intel would be the
easiest one for them to deal with.

I assume that any UDI modifications to the Linux kernel would have
merit, and not merely to support UDI.  With that and the above
stipulations, I think it is a great idea that the Linux community should
get behind and encourge Intel to make it happen!
Date: Fri, 25 Sep 1998 11:31:19 +0100
From: Charlie Stross <charlie@antipope.org>
To: lwn@eklektix.com
Subject: Thought for the day ...

Bill Gates gave a speech recently, reported in Le Monde Informatique:
he's quoted as saying that Linux will never have goodies like voice
recognition software, and that he sees it already joining OS/2 in the
history of data processing. (See Babelfish translation at

Lo, verily and forsooth, up pops the following announcement on Freshmeat
yesterday: "KVoiceControl is a speech recognition system that allows the
user to connect spoken commands to unix commands. It automagically
detects signals coming from a microphone then performs recognition on
this speech input and in case of successful recognition executes the
unix command the user hooked up to it." (See

Coming next week: NASA, NSA, IRS, CIA announce "US government has
no plans to migrate from OS/2 before 2047."

-- Charlie Stross

Date: Thu, 24 Sep 1998 12:31:09 +0100 (BST)
From: Andrew Spencer <bspajs@bath.ac.uk>
To: lwn@lwn.net
Subject: Acorn

I was interested to read your comments on the demise of Acorn's
desktop/workstation efforts. I used an Acorn machine for several years as
a teenager; I still have one at home, little used anymore because I do all
my work on an x86 Linux machine (I couldn't afford a new Acorn, and also
they didn't and don't have hardware floating-point, which I needed). 

Acorn machines, or more particularly their operating system Risc-OS,
actually have very little in common with Linux other than independence
from Microsoft. Linux is very decidedly a programmer's operating system,
in the sense that reliability and security seem to be the main things
guiding its deisgn. Acorn's operating system is quite different: much of
it was apparently written in assembler; it's based on co-operative
multi-tasking; it's built around a single-user philosophy much like
Windows, with the attendant consequences; and as far as I know does little
or nothing to stop badly-behaved programs from writing on areas of memory
that they shouldn't, or similar crimes. [Mind you, the operating system
being on a ROM -- all 2Mb of it -- maybe makes that slightly less risky.
Certainly in practice, they're better behaved than Windows machines, but
it's still not unheard of for applications to take the whole machine down 
with them.] It also doesn't really have anything to match all the Unix
tools like shell scripting, grepping etc...

On the other hand, though, no-one could describe Linux as a user's
operating system. The lack of consistency in behaviour of X applications
verges on the ridiculous, for example. By contrast, Risc-OS is really very
well designed and easy to learn: not perfect, but a model of simplicity
and consistency when compared to Windows. Some of its neatest features
are (and skip this if you don't give a toss):

*	Three button mouse: the left button acts much like the left button
in Windows, the middle button brings up pop-up menus (no menu bars) and
the right button produces clever variations on the left button.
Right-clicking on a scroll-bar tab scrolls in the reverse direction, which
is incredibly useful when scanning up and down a document. Right-clicking
a menu item leaves the menu open after performing the action, instead of
closing it. Especially good is the handling of directories. Directories
are brought up in separate windows (rather than Windows Explorer style)
i.e. double-clicking a directory folder brings up its contents in a new
window. If you double-click with the right mouse button, the window you're
currently in gets closed. If you close a directory display with the right
button, the directory above it pops up again. Much nicer than in Windows

*	Drag-and-drop: You load files by dragging them onto the
application icon (which sits on a bar at the bottom of the screen); you
save them by dragging them out of the save dialog onto a Filer window,
i.e. a directory window.

*	Applications directories: Program files and associated odds and
ends (like icon bitmaps) are kept together in a single directory per
application. This directory appears as the application icon rather than a
folder icon, and double-clicking it loads the application (it runs a
script file inside the directory called !Run). This might not be to
everyone's taste, but I like it a lot.

*	Integrated OS support for anti-aliased text since about the year
dot (well, 1988ish)...

All of which is intended to say that I'm not sure whether the demise of
Acorn's workstation division is going to lead to any significant number of
talented programmers migrating to Linux. Probably most (ex-)Acorn
programmers who might have been inclined to program for Linux already do
so. Rather, I think the message in this event is to remind us of the sad
failing of the computer industry to produce thoughtful, good and
consistent design *for the user*. We already know that technical
superiority is not, on its own, enough to win (or even survive). Ease of
use is even less so. Which is a shame, because there's more need for it
now than ever. 

LWN is excellent - well written, interesting and level headed - keep it

all the best,

Andrew Spencer

Dept of Biology and Biochemistry
 and Centre for Mathematical Biology
University of Bath
Bath BA2 7AY

From: "Zenaan Harkness" <zen@getsystems.com>
To: <editor@lwn.net>
Subject: What should we do with rms.
Date: Sat, 26 Sep 1998 09:43:42 +1000

I would briefly like to state my full support of RMS and his vision. Thank
you for printing the LWN flames that clearly expressed (I thought) the
relevance (as if that had to be stated in the first place) of RMS in _our_
Gnu/Linux world.

I would ask you not to write something so arrogantly repulsive (as the
original article) again. Do so, and I will consider not reading LWN again.

As a pennance, I suggest a change of the name of your editorial to include

On the whole I do find LWN to be an exceptional publication. I do not expect
to see such indiscrimate editorial in the future. If not so, I would expect
to see an alternative publication started. Something like "Gnu Weekly News".

Date: Mon, 28 Sep 1998 11:33:10 -0500
From: Mike Hammel <mjhammel@fastlane.net>
To: editor@lwn.net
Subject: The Linux Press:  editorial power and responsibility

I read the responses to your rms article.  A few were well worded (Danny
Yee's in particular) and argued valid points.  And perhaps they were
right in stating that your editorial should not have been out to target
Richard in the manner that it did.  But I can't help but feel that the
responders missed the opposing viewpoint as well.  The problem (yes,
problem) with RMS is not his ideas or ideals, it is in the manner in
which he expresses them.  No one likes to be degraded or verbally abused
and battered and, based only on the reports of conferences and his other
public appearances, it would appear this is this is the manner he uses
to make his point.

Of course, if this is not how he really presents his views, then those
who responded should consider taking up the journalistic trade to make
their own reports on RMS' actions.  We, people in general, make many
decisions based on the writings in the press.  The Linux press is still
young - we have limited resources and our ability to ferret out the true
story is sometimes limited by those resources.  Most of us (I speak of
writers for LWN, The Linux Gazette, Slashdot, and the various other
Linux-specific press sites on the web) are generally not professional
writers much less journalists.  We can and do make mistakes.

I've never written about RMS in my Graphics Muse column.  I tend to
avoid the issue publicly as much as possible.  But a similar editorial
about GGI generated quite a bit of unhappy mail.  Like LWN, I made a
public stance.  I wasn't quite prepared for the response.  Lesson

But its the press who have to hang themselves out on the line on issues
like these.  It gives us a power and it requires a responsibility to be
truthful and as accurate as possible.  We'll do our best.  All we ask of
our readers - all I ask - is that you not use RMS' tactics in
responding.  Be succint - we get a lot of email.  Follow up
counterarguments with verifiable proof, where possible.  And most of all
- don't be mean.  Remember, however the press may attempt to be
unbiased, it still has the power to sway the world.  You don't want to
sway them away from your point of view, do you?

Michael J Hammel
The Graphics Muse

Date: 25 Sep 1998 16:32:32 -0000
From: deivu@eccosys.com
To: editor@lwn.net
Subject: comments on the rms commentary

It looks like most of these people didn't read the original article
very closely -- though I think the author's right that it's a little

I think it's true that RMS is in danger of making himself irrelevant
(except historically, obviously), because he's an extremist -- not
just in his goals, which I think are by and large good goals, but in
his methods; he doesn't seem to see value in subversion or in
short-term compromise. The danger, of course, is losing sight of your
original goals, and perhaps that's what RMS sees happening. I don't
want to see RMS become the Leon Trotsky of the free software
revolution (but who's the Stalin in that picture?), but on the other
hand, a revolution without the support of the populace, a revolution
that confines itself to the literary salons and the debating
societies, is doomed to failure. "Open source" may be gaining
mindshare, but I don't think it's gaining it at the *expense* of "free
software"; it's gaining mindshare in places where "free software" was
too blatantly revolutionary to stand a chance. I'm sorry such places
exist, but we can't make them go away by wishing, and if "open source"
gets GPLed software into places "free software" can't take it, then
it's a victory for the Movement.

I have to disagree with the writer who says "it seems /prima facie/
obvious that Oracle and Informix ports to Linux will be good for Linux
but bad for the development of industrial-strength free database
systems." I don't think Linux ports of the large commercial databases
will have any effect on the development of industrial-strength free
database systems -- they're not going to charge significantly less for
the Linux version than for the other versions, not in the long term;
anyone who can't afford a commercial database for Solaris or NT will
also not be able to afford one for Linux.

Most computers are sold to commercial concerns and most software is
used in commercial concerns. A victory for free software has to
include victory in the business world, or it's an irrelevant sideshow
-- as long as we can't get rid of the business world altogether. My
opinion of the current issues is that anything that extends the reach
of free software in commerce is progress, even if it means short-term
compromise. The writer who says "people would be only too happy to
slowly switch to a system where they will have to pay for all but the
core components" is not thinking straight -- remember, for most
computer buyers (as opposed to coders, who look a little more closely
into these issues), that's the situation they're in *now* -- you buy
the machine, the core components come free (actually they've gone into
the cost of the machine, but they seem free), and you buy software to
run on top of those "free" components. 

Now the large commercial concerns that can afford Oracle will be able
to run it on Linux, and faced with the possibility of UNIX' stability
and security without UNIX' proprietary hardware price tag, hopefully,
many of them will choose to. They'll then have the experience of using
a free software product to do serious work, and discover that it can
in fact be powerful, reliable, and well-supported.  Today it's the OS;
perhaps tomorrow they'll decide that they don't need to pay Informix
per-user-per-processor license fees either, and move to PostgreSQL
(once it's a little faster and has a few more features) or SQRL (once
it's finished). How can this not be a good thing?


(David Moles, deivu@eccosys.com, http://www2.eccosys.com/~deivu/)


Feedback and corrections

A couple of silly mistakes slipped through last week... We mistakenly said that this C|Net Builder article was a comparison between Linux and NT 4.0; instead it compared Linux with NT 5.0 beta 2. And we don't know what slip of the fingers put the city of Madison (WI) in Ohio...

Also, Michel Emde pointed out that the "Hacker Timeline" which we had reprinted from the ISN mailing list had been taken from Bruce Sterling's "The Hacker Crackdown" book. We regret the misuse of Mr. Sterling's table and the lack of attribution.



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