See also: last week's Letters page.

Letters to the editor

From:	 tet@accucard.com
To:	 letters@lwn.net
Subject: Kernel compile times
Date:	 Fri, 15 Mar 2002 17:40:12 +0000


On the LWN kernel page, you say:

>Compiling a kernel in 23 seconds isn't bad - it looks like a record.

You then explain how Anton Blanchard rose to the challenge and achieved
a 10.3 second kernel compile. What you didn't mention was that he'd
already broken the 23s mark some 18 months ago. On 26th September 2000,
he posted a message to the Linux Kernel mailing list showing Linux
booting on a 24 CPU Sun Ultra Entrprise E10000, and claiming a kernel
compile in 20 seconds. So, a 100% increase in 18 months? Sounds like
a perfect illustration of Moore's law to me :-)

Tet

From:	 Biju Chacko <biju@exocore.com>
To:	 letters@lwn.net
Subject: Thinking with ones gonads
Date:	 14 Mar 2002 09:58:01 +0530

In lwn you wrote:

> It is worth pondering, however, on why so many of us insisted on using
> Linux systems in the early 90's, when it was still clearly inferior to
> the numerous proprietary Unix systems that were available at the time.
> Without a certain amount of "gonad thinking," Linux might not have 
> come so far so quickly.

It's fairly simple, really. In the early '90s, when I started working
with Linux, proprietary Unixes were just plain out of reach. Actually,
as a penniless student in a third-world country, DOS was out of reach! I
shudder to think of the sheer volume of piracy I commited at that time.

Why did I use Linux in '94? Not for any political reason. It was UNIX
and it was available.

Why do I use Linux in '02? Ditto. With the bonus that in many areas it
is superior to the competition.

Remember, this is a technical field we're in ... making decisions based
on politics/marketing/whatever is just asking for trouble. And *that* is
the same mistake being made by both RMS and MS.

-- Biju

PS: The .sig is a coincidence -- but appropriate nevertheless.

-- 
-------------------------------------------------
Biju Chacko        | biju@exocore.com (work)            
Exocore Consulting | biju_chacko@yahoo.com (play)
Bangalore, India   | http://www.exocore.com
-------------------------------------------------
Those who do things in a noble spirit of self-sacrifice are to be
avoided
at all costs.
		-- N. Alexander.

From:	 David Mackintosh <mackdav@rogers.com>
To:	 <letters@lwn.net>
Subject: "Gonads" vs practicality
Date:	 Thu, 14 Mar 2002 11:13:19 -0500 (EST)

Sir:

With respect to your comments on Linux's history, specifically:

> It is worth pondering, however, on why so many of us insisted on
> using Linux systems in the early '90s [...] without a certain amount
> of "gonad thinking", Linux might not have come so far so quickly.

I would doubt that any kind of religion plays a significant part of
selecting an operating system.  The vast majority of linux users have
different reasons for chosing to run this operating system.  In
general, before a user will consider selecting it for his needs, the
OS must have evolved into at least one of two states:

1.  The OS must be able to do something the user wants it to do; or

2.  The OS is in a state where the user is capable of modifying it
    to do something they want it to do.

For me, I needed a crash-resistant OS that I could afford.  Linux
fit "what I wanted it to do" and through perl, shell scripting, and
some awful awful C code, I could "make it do what I wanted it to
do" in a practical sense.

Yes, OSF/1 and SunOS were far more capable at the time, but Linux fit
my needs from an affordability perspective -- it cost $25 for a set
of CDs and ran on the less-than-state-of-the-art equiptment I owned.
I did not understand the GPL at the time, so "gonad thinking"
played absolutely no roll in the decision.

I think that the majority of the religious are involved with Linux
because using a "counter-culture" operating system is seen to be
against the mainline.  With the exception of the truely sincere (such
as Mr. Stallman) I think these people are useless as a core
constituency, because they will flee Linux and the GPL as soon as it
becomes mainstream in favor of something more "revolutionary".
Perhaps this is where the Hurd's userbase will come from?

-- 
 /\oo/\
/ /()\ \ David Mackintosh | mackdav@rogers.com

From:	 Joe Klemmer <klemmerj@webtrek.com>
To:	 letters@lwn.net
Subject: "gonad thinking"
Date:	 Thu, 14 Mar 2002 16:38:09 -0500 (EST)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


> It is worth pondering, however, on why so many of us insisted on using
> Linux systems in the early 90's, when it was still clearly inferior to
> the numerous proprietary Unix systems that were available at the time.  
> Without a certain amount of "gonad thinking," Linux might not have come
> so far so quickly.

	FWIW, I think there's a difference of scale and/or perspective.  
When Linux came out there was no expectation of it being or doing real
work.  It wasn't needed to, nor desired for, something that actually did
anything.  BitKeeper, OTOH, is expected to perform a function that has
some requirement for someone (specifically LKH'ers).

	But that's just my opinion.  What do I know... 

- ---
Using Linux since 11/91		|		http://www.linux.org
Linux user #29402		|		http://counter.li.org/
Red Hat Linux			|		http://www.redhat.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8kRhGHeWRPx8OIHARAn2WAJ9ACa8cWyd5Kh6XQb0fUnKW2REDhgCdHF+x
6Vuw/7xOvXAwwD6Ps1HXZ94=
=+tra
-----END PGP SIGNATURE-----

From:	 Zygo Blaxell <zblaxell@feedme.hungrycats.org>
To:	 letters@lwn.net
Subject: Whoa there!  Linux inferior to _which_ Unix systems in the early 90's?
Date:	 Fri, 15 Mar 2002 23:22:04 -0500

>It is worth pondering, however, on why so many of us insisted on using
>Linux systems in the early 90's, when it was still clearly inferior
>to the numerous proprietary Unix systems that were available at the
>time. Without a certain amount of "gonad thinking," Linux might not have
>come so far so quickly.

Whoa there...I have a few problems with that statement!

Last time I checked, I was a Linux and proprietary Unix user in the
early 90's.  I was doing a fair amount of "gonad thinking" at that time
in my life, but I certainly wasn't thinking about _software_ with those
organs.  Uhhh...  maybe that's more than my fellow readers want
to know.  Anyway...

First of all, the world is not all Unix--by 1993, there were many
non-Unix operating systems for 386-class machines that were clearly
inferior to Linux.

Even if we compare Linux only to proprietary Unixes, I'm having difficulty
recalling which proprietary Unix vendor had a product that Linux was
"clearly inferior" to at the time.  Each proprietary Unix vendor optimized
their product for one (usually vertically integrated) niche market,
and utterly ignored anyone else, so each proprietary Unix was good
at one particular kind of application and bad at most others.  

Linux was not "clearly inferior" to contemporary Sun, HP, IBM, SGI,
or SCO systems of the early 90's--I used a bit of all of those at
the time, as well as Linux, and they all sucked more or less equally,
especially if you wanted to run a mixture of many different applications.
This lack of generality was part of the very reason why there were so
many proprietary Unix vendors in the first place!

People who were serious about working on a Unix system (as opposed to
merely running a single application on it) in the early 90's routinely
replaced as many of the bundled system components with third-party
replacements as possible, using proprietary software if you had money, or
GNU and other free (beer) software if you didn't.  The major motivation
for this was to have a layer of portable runtime to insulate you from
the vendor's Unix taste, so you could develop software for multiple
vendor platforms without going insane.

By the mid 90's, there was very little practical difference between
a proprietary Unix system that had been mostly hidden under a thick
layer of third-party components, and a Linux system that didn't have
any vendor-supplied components to start with.

Oh, and a lot of this third-party software was free (speech) too.
Wasn't that nice!

-- 
Zygo Blaxell (Laptop) <zblaxell@feedme.hungrycats.org>
GPG = D13D 6651 F446 9787 600B AD1E CCF3 6F93 2823 44AD

From:	 Eric Smith <eric@brouhaha.com>
To:	 letters@lwn.net
Subject: nore GPL confusion
Date:	 20 Mar 2002 22:20:45 -0000

Gentlemen,

In your 20-Mar-2002 daily updates, you quote a Financial Times article
as saying "the GPL can 'convert' proprietary software into open source
software - since any company incorporating licensed code into its own
software products is obliged to open up its code too."  You do state
that this worry is "overblown", and I agree.

Developers should remember that using GPL'd code is a choice that is
available to them, not a burden.  If the terms of the GPL are not
consistent with their intention to keep their software proprietary,
they should not use the GPL'd code.  No one is forcing them to do so.
The have the OPTION of using GPL'd code IF they desire to AND they
are willing to comply with the license.

The way the Financial Times article was worded, it almost sounds like
they think GPL'd code can somehow mysteriously sneak into another software
package when the developers aren't looking.  Unless the software developers
are grossly incompetent, that's not going to happen.  It is a legitimate
concern that software developers must be aware of the licensing on any
software that they merge into a product, but this is in no way specific
to the GPL.

Sincerely,
Eric Smith

From:    Leandro Guimarães Faria Corsetti Dutra <leandrod@mac.com>
To:	 letters@lwn.net
Subject: The GNU HURD will be ready by the end of the year
Date:	 Thu, 14 Mar 2002 14:30:04 +0100

	First, you've taken RMS words off a much edited interview... not good 
for acuracy.

	Second, you got it all right about keeping proprietary software off 
the GNU system.  Obviously RMS isn't talking about applications 
being able or not to run.  He's talking about "Distributions of 
GNU/Linux", and obviously FSF's distribution of GNU, being based on 
Debian, won't include any proprietary software.

	About licensing, obviously GNU GPL libraries won't allow proprietary 
software.  This is already the case with GNU libraries like readline 
and some Gnome ones, as well as with Troll Tech's Qt -- you have to 
get a different license to be able to distribute your application 
with Qt.

	On the other hand, applications usually don't talk to the kernel, 
they talk to libraries such as GNU libc, and most of these have been 
licensed under LGPL specifically to allow for proprietary applications.


-- 
 _
/ \ Leandro Guimarães Faria Corsetti Dutra        +41 (21) 216 15 93
\ / http://homepage.mac.com./leandrod/        fax +41 (21) 216 19 04
 X  http://tutoriald.sf.net./               Orange Communications CH
/ \ Campanha fita ASCII, contra correio HTML      +41 (21) 644 23 01

From:	 Jeroen Dekkers <jeroen@dekkers.cx>
To:	 letters@lwn.net
Subject: The GNU Hurd
Date:	 Tue, 19 Mar 2002 17:52:33 +0100

It's nice that you write something about the Hurd, but it would be
better if you would have looked at the Hurd homepage,
http://hurd.gnu.org, a bit better. To start with, you write it as "the
Hurd", not capitalized like "HURD". See
http://www.gnu.org/software/hurd/faq.en.html#q1-2 for more
information.

Second, the Hurd isn't a kernel. It's a multi-server system running on
a microkernel, Mach at the moment. In the future the Hurd will
probably run on L4 (http://os.inf.tu-dresden.de/L4/ and
http://www.l4ka.org) and other microkernels.

There are plenty of reasons for producing the GNU system. One of the
reasons is to give attention to the GNU project. GNU/Linux never did
(at least not before RMS started with his GNU/Linux campaign) give
much credit to GNU. (Just name LWN for example. Your article about the
Hurd perfectly shows that it's misnamed). The GNU system will do. It
also generates money if the FSF is going to sell cds with the GNU
system. Money which will be spend to advocate and develop free
software.

You really misinterpretted RMS his reasons for creating the GNU
system. He only says that he's looking forward to see the GNU system
available because it will only contain free software. He didn't say
anything about forbidding the user to install non-free software on it
if that user doesn't care about his freedom.

The following sentence doesn't make much sense to me either: "Thus, it
seems unlikely that the HURD will mount a substantial challenge to the
established free kernels anytime soon." And that's not only because it
doesn't make much sense seeing it stand-alone, the potential of the
Hurd is very big, but because the arguments before it. 

Why shouldn't it be able to challenge the existing free kernels
because the GNU distribution doesn't contain non-free software? I'm
not sure if you know that one of the most populair distributions,
Debian, doesn't have non-free software in its main
distribution. Debian works perfect without the non-free section and a
lot of people I know have are running it without it. Proprietary
software isn't necessary any longer. Debian also has a Hurd 'port',
BTW.

But what is actually the biggest reason? It's because of its technical
superiority. Although the current implementation doesn't show it, the
design of the Hurd and the ideas behind it really rock. And of course
I get flamed about this by people who think microkernels suck. I'll
just say that research proved that it's possible to make a good system
based on a microkernel. It can be fast, for example. That there isn't
a good system at the moment doesn't make it impossible. All flames
based on FUD will go directly to /dev/null.

The Hurd has many nice features. To give an example, you can run
servers implementing file systems, network protocols, file access
control, etc just as a normal user and debug it as a normal
program. Any user can add things to the filesystem, for example
mounting a directory on a ftp server in its home directory. All this
is possible without special permissions.

And all those things and more are possible because the Hurd is a
multi-server system running on a microkernel. This is the reason why
the Hurd is still developed and why the number of Hurd developers and
users is increasing.

It's nice that the Hurd is getting more usable. You would probably ask
why the Hurd taked so long. The answer is simple: manpower. The Hurd
only has a few developers and never had much in the past. The Hurd is
a redesign of Unix, a redesign takes much more time than a
reimplemention. That's why Linux was much easier to develop, they
didn't have to think about the interfaces and the design.

But one thing is sure: The Hurd isn't dead and the GNU system has a
lot of potential to beat a lot of operating systems, even those much
used proprietary ones.

Jeroen Dekkers
-- 
Jabber supporter - http://www.jabber.org Jabber ID: jdekkers@jabber.org
Debian GNU supporter - http://www.debian.org http://www.gnu.org
IRC: jeroen@openprojects

From:	 "Robert A. Knop Jr." <rknop@pobox.com>
To:	 letters@lwn.net
Subject: reverse FUD
Date:	 Fri, 15 Mar 2002 13:13:11 -0600

It's striking that as the news of the zlib flaw gets out, what we're
seeing is lots of patches from Linux and FreeBSD distributors, and at
the same time realization by Microsoft that maybe this is their problem
too.  Yes, this flaw is a bit of a PR black mark for open source, but it
is interesting to note the speed of response of the two different
communities.

There's another take on this too.  Assuredly this security flaw is going
to be a giant pain for Microsoft, if they have to patch all sorts of
applications as a result.  BUT, if zlib had been released under the GPL,
Microsoft wouldn't have used it, and they wouldn't have this problem!
So... Microsoft likes to go around saying that GPL "breaks the cycle"
that they think ought to be present in the computer industry.  But,
really-- when somebody releases something under the GPL, they're just
trying to protect you, Microsoft, so you won't be succeptible to its
flaws!  It's all done for your own good.  Really.

-Rob

-- 
-=-=-= Rob Knop =-= rknop@pobox.com =-= http://www.pobox.com/~rknop =-=-=-
     Help the EFF protect basic freedoms online: http://www.eff.org
 Playwrights & theatre types, see The Dramatic Exchange: http://www.dramex.org

From:	 "David L. Craig" <dlc@radix.net>
To:	 letters@lwn.net
Subject: On Overstatement
Date:	 Fri, 15 Mar 2002 16:37:18 -0500

In "Cal Senator: Hollywood Over Tech (Wired)," the abstract
ends with "Yes, it's true: the U.S. government really wants
to outlaw free software."  Now, honestly, don't you think
that is retractable nonsense?  It is proper to attribute
this to some members of Congress, but to the whole kit and
kaboodle?  These new bills are still in committee, after
all.  If you remember the U.S. electorate IS the U.S.
government, you will see how overstated this was.  In fact,
I perceive it to be inflamatory and beneath the journalistic
standards I have come to expect from your Web site.  I beg
you, fix this pronto.

-- 

May the LORD God bless you abundantly!

Dave Craig

-  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -
"So the universe is not quite as you thought it was.
 You'd better rearrange your beliefs, then.
 Because you certainly can't rearrange the universe."

--Athor 77, formulator of the       from _Nightfall_
  Universal Theory of Gravitation   by Asimov/Silverberg