See also: last week's Letters page.

Letters to the editor

From:	 Dominic Mitchell <dom@semantico.com>
To:	 letters@lwn.net
Subject: IPFilter
Date:	 Thu, 24 May 2001 17:44:01 +0100

It's worth noting that (at least in the case of FreeBSD, not sure about
the others), there is also ipfw, another kernel level firewall, which
is competent to work with, if you do not like ipfilter.

This is similiar to the situation with floating point emulator in
FreeBSD.  There are two distributed versions, one GPL, one BSD.  You can
choose the one that applies.

-Dom

From:	 "Chad C. Walstrom" <chewie@wookimus.net>
To:	 letters@lwn.net
Subject: The Boundaries of GPL
Date:	 Tue, 29 May 2001 11:29:20 -0500

You wrote:

    But the real problem is that Linus does not own the copyright for
    the entire kernel. Many major contributors have retained their own
    copyright on the code they have added, and many of them are
    opposed to proprietary modules. That leads to a couple of
    troublesome scenarios:

An interesting possibility to solving this problem is to unify the
copyrights of all code contributed to the Linux kernel.  There are two
ways to do this, transfer copyright control to the Free Software
Foundation, or create a non-profit organization specifically designed
to maintain and develop the Linux kernel and all aspects thereof.  I
highly doubt that all the Linux kernel developers could be convinced
to sign over copyright control to their contributions to the FSF, as
not too many people buy in to the Marxist-like views of RMS and the
FSF.  

The second option is quite intriguing, however.  An organization that
is focused strictly upon the Linux kernel would guarantee that the
concerns and issues related to the kernel would not be overlooked in a
larger, all-encompassing organization.

As a side note, I have been unable to find any documentation
concerning an established Constitution of the FSF or any such related
issues.  How is the FSF run?  Who is in charge?  Who defines policies?
Is it an RMS monarchy?  The GNU website is only helpful in so far as
providing the GNU Manifesto and multiple opinion papers on Free
Software.

-- 
Chad Walstrom <chewie@wookimus.net>                 | a.k.a. ^chewie
http://www.wookimus.net/                            | s.k.a. gunnarr
Key fingerprint = B4AB D627 9CBD 687E 7A31  1950 0CC7 0B18 206C 5AFD

From:	 Neilen Marais <nmarais@hertz.ee.sun.ac.za>
To:	 floydls@home.com, letters@lwn.net
Subject: On The Desktop May 10 -- KDE Bloat
Date:	 Thu, 24 May 2001 09:51:17 +0200

Hi Floyd, LWN people.

> Also, I agree completely with Michael A. Schwarz in his email (Wrong 
> way to look at it).  The time to make a program fast and use less 
> memory is when it is designed and implemented, not later.  If you 
> wait till later then you miss the most beneficial time to improve 
> it.  KDE and GNOME should work OK on old equipment.  After all MS 
> windows works and KDE and GNOME.  They are not all that much more 
> advanced.

While I agree with the fact that you'd be missing the best time to
optimize for speed if you do it later (i.e.. later people will have
faster hardware anyway), it is just not always (seldom, really)
practical to make things fast from the start.  

In the words of Donald Knuth (whom you should all know of :-)

"Premature optimization is the root of all evil".

While a good initial design would make it easier to improve performance
later, it is unlikely that early implementations of any design will have
optimal performance.  

In the case of desktop environments there was virtually nothing to start
with, and thus features tend for the moment to take preference to
speed.  IMHO this is the right approach to take.   Says even I running a
4 year old low-end system... But upgraded to 160 MB RAM. RAM is _cheap_
right now.

Cheers
Neilen

P.S. I have no specific involvement in any of the desktop projects, but
remember how excited I was to find the very very first KDE activity.

From:	 Eric Smith <eric@brouhaha.com>
To:	 letters@lwn.net
Subject: ioctl() replacement
Date:	 24 May 2001 18:07:40 -0000

In your 24-May-2001 issue, you mention one of the possible ioctl()
replacements:

    Another approach calls for the opening of a control channel as a
    separate file descriptor, then invoking operations with write() and
    read() calls. Such an approach is workable and network-friendly, but
    it lacks the atomic nature of ioctl(). Things can happen between
    when an operation request is written and the results are read back.

The latter objection is not necessarily true.  If the interface is defined
such that a response is always generated when the write() is performed, as
part of the atomic operation, and the response is simply not delivered
until the read() call occurs, I don't think there's a problem.  At least,
not any new problem that didn't already exist with ioctl().

Being network-friendly is a huge benefit.  I have needed for some time
to be able to fully control remote serial ports and remote tape drives.
I thought about doing a hack to work like ptys but for arbitrary devices
(rather than just ttys), and somehow bouncing ioctl()s to user space, where
a process would act as a proxy to get the calls to the remote machine.
I quickly learned how infeasible that is.  Linus' proposal to change the
ioctl() call interface, while not directly network-friendly, at least
would make it possible to support my specialized proxy.

Eric Smith

From:	 Martin <chimbis@skjoldebrand.org>
To:	 lwn@lwn.net
Subject: Linux on desktops
Date:	 Fri, 25 May 2001 07:45:23 +0200

I found I couldn't agree with the DELL representative quoted as:

"It's still a fundamentally technical operating system," said Steve 
Smith, Dell's European market development manager for client systems. 
"It's very easy for someone who doesn't know what they're doing to 
break. It's not designed for the novice user." 
Microsoft's grip on the applications market, with Office, also shows no 
sign of weakening, Smith said."

This might reasonably make sense only from a narrow point of view - 
namely people and countries that actually can afford throwing away 
thousands on tools to communicate. Many people do not have this kind of 
money available for something that isn't necessary for everyday 
subsistance. It is surprising that a highly configurable, stable, 
useable and affordable philosophy isn't being actively "marketed" 
towards people and countries where cost is a big issue. Specifically 
developing countries could benefit imensly from using an operating 
system, and applications easily and cheaply customized for specific needs.
My plan is to try to (together with organisations that has expressed an 
interest for this in our standard aid program) implement a GNU/Linux 
based computer training center (or even centers) in East Africa during 
2002.

I'm interested in hearing from anyone who has been involved in similar 
projects in the past. You're welcome to write to me privately 
(chimbis@bahnhof.se) or at my office address 
(martin.skjoldebrand@forumsyd.se).

Cheers,

Martin Skjöldebrand
CTO, Forum Syd
Swedish NGO Center for Development Co-operation

--
Speaking only for myself. Views expressed in this email message are not 
necessarily shared by my employer.

From:	 Bret Mogilefsky <mogul@gelatinous.com>
To:	 letters@lwn.net
Subject: Letters page for May 24th
Date:	 Thu, 24 May 2001 16:28:57 -0700

Rob Funk wrote in your letters page:

  While I generally respect Hammel's work, like others I am beginning to
  question how his writing fits into LWN.  It reads more like a column than
  a section of news, and none of the other sections read like that.  [...]
  I believe that desktop-oriented news is important, but the current style
  doesn't inspire confidence in this news source.

I agree that desktop news should be well covered, and that the current
writer is just inappropriate.  It is quite frustrating to watch him bumble
through learning what's available and how it works week after week as
he's writing rather than writing about new developments or issues.  He's
confessed several times that he's new to the whole desktop thing and
started writing copy about Gnome before he'd even tried KDE... 

I consider LWN to be the highest quality Linux news source available, and
in my mind it needs knowledgable writers the way ABC World News Tonight
needs knowledgable reporters.  If we watched a reporter on Capitol Hill try
to explain what he does and doesn't know about Congress at the same time
he's reporting on what legislative issues were being hotly debated that
day, we'd probably change the channel... As it is now, I can only go two or
three paragraphs into the section before I skip to the next page in
frustration.  This writer is better suited to writing something like
"Desktop tip-of-the-week" than writing the insightful, comprehensive weekly
column we expect.

Sorry, but he's got to go...


Bret

From:	 =?iso-8859-1?Q?J=F6rn?= Nettingsmeier <nettings@folkwang-hochschule.de>
To:	 letters@lwn.net
Subject: The "Desktop" section needs rethinking.
Date:	 Sat, 26 May 2001 13:22:59 +0200

Dear LWN editors !

I've been reading lwn.net for about 2 years, and i find it a very
high-quality and concise news source. I greatly appreciate your work
and the continuity of a weekly up-to-date newsletter.
The "Desktop" section, however, leaves a lot to be desired and does
not at all match the level of the other columns.
The irritating verbosity there (as opposed to the dense and
informative style elsewhere) has been pointed out by others before. 
What actually triggered this little rant of mine was the author's
question of how to remove desktop icons. While he claims to be an
old-school Fvwm user (which is perfectly ok), my very humble opinion
is that an editor of "On the Desktop" should be familar with
concepts such as right-clicking for a context menu. Instead of
providing insightful news about the latest developments, the section
reads more like the very first steps of a dedicated bare-X user on a
modern GUI system. While this is certainly interesting, it does not
meet the expectations I have of a "Desktop" column in LWN.

No offense intended,

sincerely yours,

Jörn Nettingsmeier



-- 
home://Kurfürstenstr.49.45138.Essen.Germany      
phone://+49.201.491621
http://icem-www.folkwang-hochschule.de/~nettings/
http://www.linuxdj.com/audio/lad/

From:	 Pavel Roskin <proski@gnu.org>
To:	 <letters@lwn.net>
Subject: Is Nautilus defunct?
Date:	 Fri, 25 May 2001 10:15:36 -0400 (EDT)

>From the daily news for May 25:

> Still, they offer many installation options ranging from a complete
> text-based Linux system up to full GUI splendor with the latest
> version of KDE or GNOME equipped with Eazel's now defunct Nautilus.

I believe that "defunct" is a wrong word for Nautilus. Let's see how
Webster dictionary defines this word:

Etymology: Latin defunctus, from past participle of defungi to finish,
die, from de- + fungi to perform -- more at FUNCTION
Date: 1599
: no longer living, existing, or functioning <the committee is now
defunct> synonym see DEAD

Nautilus is living, existing, and functioning, just as it was before.
Free software doesn't die if the principal developers stop working on it -
it dies when it's no longer needed.

-- 
Regards,
Pavel Roskin

From:	 Joe Klemmer <klemmerj@webtrek.com>
To:	 letters@lwn.net
Subject: Automatic updating
Date:	 Fri, 25 May 2001 11:15:59 -0400

On the subject of tools for automatic updating of software you said...

"One other note on automated installers: the Debian project has
probably had this issue addressed longer than anyone with their set
of "apt" tools. Use of apt to update just about any package is pretty
painless. All that may be missing from this might be a graphical
front end integrated into KDE and GNOME."

The Free/Net/OpenBSD way of doing updates is very much like the apt
tools (from what I understand).  Granted I haven't actually used either
of these but I know that on the *BSD's you just go to the install
directory and do "make install" and it'll go out, find all the software
and dependencies, build and install everything.  This seems to be, at a
fundamental level, the same way that apt-get works.

I mention all of this because, if memory serves (and that would require
an act of god), FreeBSD was using this system before Debian was
created.  As I said, I could be way off on the timing, but it would be
good to see just how similar or different that the Debian and *BSD ways
really are.  I've been using RPM's for, well, since RPP evolved into
RPM.  However, I really like the idea of just doing something like -

# cd /usr/src/ports/xfce
# make install ; clean

and having my desktop all updated.

Joe

---
"Do not meddle in the affairs of wizards, for you are crunchy and taste
good with ketchup."

From:	 Ronald Pottol <chaos@chaostrophy.org>
To:	 letters@lwn.net
Subject: AMANDA
Date:	 25 May 2001 15:23:03 -0700

> > From the May 17,2001 LWN:
> > 
> > "That said, it is worth pointing out that, as far as we know, there is
> > still not a free, top-quality large network backup and restore system
> > available for Linux. Numerous commercial alternatives are out there, but
> > the available free systems just do not have the same level of features and
> > scalability. This could be a good project for somebody..."
> 
> Perhaps you missed a nice package by the name of AMANDA
> (http://amanda.org/).

AMANDA is a neat package, but it has two flaws that I see, first, it
cannot split a backup (single partion) across a tape (and it only
supports tape), second, there is no easy way to get a minimum set of
tapes for an off site backup, as it spreads your full backups across
your tapes.  Handy in many ways, but ugly if you want a few tapes that
have a full backup of your site on them for an off site backup.

Add these two things, and I think it would really be there.  

Ron
-- 
All your CPU are belong to Tux!

From:	 Max.Hyre@cardiopulmonarycorp.com
To:	 thanks@amazon.com
Subject: Why I'm not availing myself of ``Great Savings Every Day''
Date:	 Tue, 29 May 2001 11:39:10 -0400
Cc:	 ask_tim@oreilly.com, letters@lwn.net

[Note:  This is intended for Mr. Alan Brown, or someone with a
 similar level of authority.  The reason should be obvious once
 you've read the body.]

   Dear Mr. Brown:

   Under the subject line of ``Get Great Savings Every Day''
you wrote to me:

> We noticed it's been a while since you last ordered from us
> (using this account).
	[snip]
> Sincerely,
> 
> Alan Brown
> Grand High Pooh-Bah of Fun and Convenience
> (a.k.a. Chief Marketing Officer)
> Amazon.com


   Actually, I've already written to your company about why
I'm no longer a customer, but from the tenor of the response,
it was handled by someone with no authority to either address
the problem, or refer the matter up to someone who might.

   In short, I am sending my trade elsewhere until Amazon
ceases to support software patents.  That's why I returned my
last purchase from you, and have made none since.

   ``But, we don't'', I hear you say, ``we're just forced into
using them by competitive pressures.''  This is right up
there with any rationalization of the form ``We have to <X>
in order to compete fairly with others who do'', where you
can fill in <X> with whatever is the standard ethical lapse
for the activity in question.  Pick whichever of `bribe',
`smuggle', `deal with a corrupt regime', `write misleading ad
copy', etc., etc., as seems to you to match software patents
for odiousness.

   As a computer programmer, my ability to practice my craft
is directly threatened by software patents.  Software is a
malleable medium, probably the ultimate such.  Any programmer
worth her salt is continually inventing new forms, discarding
those inappropriate to the task at hand, shaping others to
fit the current need.  Such work requires skill and undivided
focus to execute even halfway proficiently.  Such conditions
are impossible under a regime which requires the programmer
to continually look over her shoulder for fear the patent
police are coming to say she can't use this module, that
function, the other line of code.

  I am further taken aback that a company whose lifeblood is
in the ebb and flow of computer instructions, would counte-
nance such a shackling of the employees who create and oversee
that ferment.  If this is not apparent to you (and I must
conclude that it isn't), I urge you to talk to your program-
ming employees, taking with you an assortment of the software
patents granted in the U.S.  Ask them how they can work
without violating one or the other of them.  Ask them what it
would take to know every patent issued, including yesterday's
batch, in order to dodge such violations.  If they're honest,
they'll tell you it's a mental impossibility.  The software
patent will ultimately drag to a halt the most fruitful field
of endeavor seen in the history of technology, all in the
name of fostering advances in the state of the art.

   Where were software patents from, say, 1945 to 1970?
Nowhere to be seen, as computing started the exponential
growth that continues today.  They are not a necessary
ingredient for computing's growth, but they are sufficient
to stunt, if not halt, it.

   Thus, I ask you, Mr. Bezos, and the board of directors
of Amazon to examine this situation in the clear light of
day, and join in the fight to abolish software patents
entirely.  It's required, O Pooh-Bah, to restore the
``Fun and Convenience'' to programming.


		Sincerely,

			your ob'd't servant,


						Max Hyre