See also: last week's Letters page.

Letters to the editor

From:	 Joey Hess <joeyh@debian.org>
To:	 letters@lwn.net
Subject: lsb, debian, &etc
Date:	 Wed, 4 Jul 2001 12:07:06 -0400

First, a minor correction: The LSB does specify the locaton of init
scripts -- they must go in /etc/init.d/:

	An init.d file is installed by copying it into /etc/init.d
	(which may be a symlink to another location).

It's rather disappointing that you characterize Debian's response to the
LSB as mere "grumbling". Yes, there has been some grumbling (some of it
from Debian developers who tried to participate in the LSB in the past
and feel their contributions were rebuffed and ignored).

But we have also pointed out several holes in the LSB's specification of
the rpm subset the LSB specifies. Some of these holes, unless closed,
could well make alien _not_ be sufficient to fully support LSB packages
on Debian. We've pointed out other problems in the LSB that are
unrelated to the whole RPM issue. I am hopeful that the LSB recognizes
the value of constrictive criticism, even though LWN chooses to
characterize it as "late and unfounded", and that the LSB will resolve
these problems now that we have brought them to their attention.

In the meantime, without a clear spec, I can't modify alien to support
LSB packages, and it seems that Debian cannot commit to supporting the
LSB in the near term.

-- 
see shy jo, speaking as the author of alien, and a Debian developer, but
            not speaking for Debian as a whole

From:	 Jan <jandersen@striva.com>
To:	 letters@lwn.net
Subject: The challenge
Date:	 Wed, 04 Jul 2001 11:43:55 +0100

In the article 'A challenge for the free software community' it is stated
that the freeware community isn't innovative and that we need to prove
ourselves. Well, perhaps. I'm not so sure about that - what is more likely
to prevail is reliability, durability, credibility etc etc. Hasn't it been
shallow 'innovations' that have ridden Windows like a nightmare all these
years?

The true innovation in Linux and freeware lies in the revolutionary concept
of doing something properly and for free. True innovation doesn't just
happen because we try hard; in fact it happens more often because we DON'T
try, in the process of finding a good solution to a real problem.

As for the passport thing - whose problem is it that is being solved here?
Yours and mine? Do people really want this? I don't say that it couldn't be
useful or that it won't become popular, but what would people really want,
if they could have it entirely their way? MS Password hasn't been made to
help users - it is a device similar to the many 'loyalty cards' and
whatever, that you get in most superstores. The intention with these
concepts is not so much to hold on to the customers, but to analyse their
spending habits, so that the shops are better able to manipulate people
with 'targeted marketing'. Because, as I think everybody knows, traditional
marketing isn't efficient enough - but that's another point altogether.

If we need to innovate, then let's innovate deeply. Freeware has always
been 'of the people, for the people' - so let's introduce something that
really solves actual problems for people rather than the big businesses.

How about starting with our basic values: like free sharing of knowledge
and sustainability. I think everybody knows that although the big,
predatory, 'growth oriented' businesses seem spectacular, there's only room
for a limited number of them; the real backbone of any society is the
working people and the businesses, mostly small, that are satisfied with
earning a good enough living.

In my opinion freeware isn't really about writing software and distributing
it under a certain type of license - it is a new attitude or mindset.  What
we are doing is a revolution, not violently, but simply by presenting the
world with something that is obviously right. So you could say that this is
political as much as, or even more than anything else; and this mindset of
ours has an overwhelming strength, simply because it is 'good' or 'true' or
whichever word you want to use. That is why Microsoft fear the freeware
movement more than anything else; because in the end they simply can't win,
there's just no way.

/jan

From:	 dps@io.stargate.co.uk (Duncan Simpson)
To:	 letters@lwn.net
Subject: Is .NET as threat?
Date:	 Sat, 14 Jul 2001 20:35:20 +0100

Does anyone know what .NET is yet? AFAIK nobody has any technical information,
except perhaps that everyone is expected to pay M$ on a regular basis,
including developers. If we, the free software community, can create a model
where people get the same results without this tax then that could be
compelling. I do not think we need .gnu or anything similar to .NET to do
this: instead  a sufficient set of, probably a reletvively small number of
reletively simple bits of software, should suffice.

M$ needs something like .NET for long term revenue and the fact that their
business model requires people to regualr pay for software. IF this is not
part of your business model then you do not need most of the obvious bits of
.NET. As for the secure communiaction software much of it exists alreadu
in particular their is ssh, openssl, etc. Free software not requiring a
permanent network connection could be compelling in places where people pay
quite a lot per minute for network connectivity (which only does 56k in and
33.6k out).

Duncan (-:

From:	 "Matthew B. Kennedy" <matthewbk@yahoo.com>
To:	 letters@lwn.net
Subject: A challenge for the free software community
Date:	 Fri, 6 Jul 2001 14:37:11 -0700 (PDT)

I'm always surprised to read about the need to build an Open Source or Free
Software equivalent .NET. I thought the free software community already had
dozens of such contenders? .NET is blatant catch-up technology meant to compete
with (J2EE) Java Enterprise technologies (a set of open specifications Sun
developed -- http://java.sun.com/j2ee).

There are many Free Software or Open Source implementations of the J2EE
standards. There are several J2EE application servers -- JBoss
(http://www.jboss.org) is a __truly_marvellous__ GPL'd EJB server/container.
Several JSP/Servlet servers: Jetty (http://jetty.mortbay.com) and Tomcat
(http://jakarta.apache.org/tomcat) are both Open Source. You can't poke a stick
at the number of excellent XML tools for Java. 

There are many more I can't mention here.

These tools all measure up to what .NET essentially provides. How about the
additional points from that article? Well:

   *  The community needs to design a framework which handles tasks like
      authentication and transactions

^^^ This is core to J2EE

   *  The full set of protocols which implement this framework must be
      open, with an open development and extension process.

^^^ Provided by the open set of J2EE specs.

   *  No one company or institution should be indispensable to the
      operation of the framework. No company or institution should be
      able to dictate the terms under which anybody may participate in
      life on the net.

^^^ None of the plethora of commercial J2EE vendors have twisted the specs yet
- and it wouldn't be in their best interest to do so.

   *  Security and privacy must be central to the framework's design.
      All security protocols must be open and heavily reviewed.

^^^ Such is already the case on the J2EE platform.

   *  The framework must bring the net toward its potential as the
      ultimate communication channel between people worldwide, and it
      must allow the creation of amazing new services and resources that
      we can not yet imagine

^^^ I am sure this is possible on the J2EE platform :-)

So you see, it really is puzzling to me why we need to create a .GNU when we
already have an impressive and (I think) superior Free Software and Open Source arsenal.

From:	 "Jonathan Day" <jd9812@my-deja.com>
To:	 letters@lwn.net
Subject: .NET, Passport and Bugbears, oh my!
Date:	 Thu, 5 Jul 2001 05:21:14 -0700

Dear editors,

  Let's start with examining what these programs do. As you say, they
provide mechanisms for authentication and secure(-ish) transactions.

  The question is not "how do we do these things under Linux", as many of
us already do. If it were simply a matter of replicating Passport and .NET,
the battle would already be won. Kerberos 5.2, OpenSSH, OpenCA, the
International Patch and FreeS/WAN's IPSec provide all these tools, and much
more, besides.

  (Kerberos gives you your authentication, at the user level. OpenSSH then
gives you strong encryption for transactions. The International Patch, plus
IPSec, allows you to safely authenticate a machine. Finally, OpenCA gives
you a means to roll certificates that contain any additional authentication
you may want, beyond that which you already have.)

  The problem is, Joe Average doesn't have time to decrypt the manuals for
these, let alone plough through obscure command-line interfaces to actually
get anything done. Besides, once they -have- got something done, can you
name any servers which accept Kerberos tickets from remote Kerberos systems
for authentication?

  The solution seems simple enough. The bricks exist, the cement exists,
and there are plenty of sample GUI interfaces which can serve as
architects. This just leaves the builder. There doesn't seem to be one with
an itch strong enough.

  In summary: The "Free Software" and "Open Source" communities already
have software that can blow the socks off Passport and .NET. It's proven,
reliable, and well-tested. It's just not used. Fix that, and you've fixed
the .NET for good.

Jonathan

From:	 "Robert A. Knop Jr." <rknop@pobox.com>
To:	 <letters@lwn.net>
Subject: What I want to know about DotGNU and GNU Mono
Date:	 Mon, 9 Jul 2001 07:22:10 -0700 (PDT)

All of this noise about Free Software replacements to .NET is, with some
hesitation, encouraging.

I hesitate because I've only read the public annoucements out of Ximian
and the FSF.  From those, it sure looks like what we've got here is a
last-ditch catch-up response to Microsoft's .NET initiative-- not the bold
grabbing of the reins and taking the lead that LWN.net suggested was
necessary a week or two ago.  The very name is suggestive... "DotGNU".
Maybe it's necessary for people to realize what this is, but (a) it makes
the Free Software community look like it's trying to provide a (doubtless
to be perceived as poor) substitute, and (b) it probably opens up the
whole project to the sort of "trademark infringement" attack that the
vector drawing program in KOffice was subject to.

If .NET is going to come, and we're going to have to deal with it, I for
one sure do want there to be a Free Software solution that lets us deal
with it without being dependent at all on Microsoft.  For this reason, I
will support the projects (and might even, given time, see if I can
contribute to them).  I currently live in fear that the great world we
have today (where Linux *is* a viable alternative) is going to be gone in
a few years, thanks to Microsoft's takeover of the internet under the
guise of .NET.  Anything to stop that is good.

We are, however, seeing a fundamental change of balance.  The internet was
built on open standards;  Microsoft got to the game late, tried to take
over as it took over the desktop maket, and at first simply didn't
succeed.  Nobody was interested in eschewing the Internet for MSN.
Later, Microsoft's servers couldn't push the Unix (and eventually Linux)
servers out of the market, despite Microsoft's dominance on the desktop...
and for this reason, Microsoft had to continue to conform to the open
standards of the internet.  Microsoft was playing catch-up.  But Microsoft
has finally realized that the age old adage applies: "If you can't beat
'em, take 'em over."  No longer, does it seem, that the internet will be
built on open standards which require commercial companies to adapt.  No,
now Microsoft is specifying how the new parts of the internet will work,
and the Free Software community will have to struggle to provide
non-Microsoft implementions of this Microsoft-specified internet.  This to
me is extremely and fundamentally sad.  What is the purpose of all the
inroads Linux has made on desktop functionality, if now all of a sudden in
the internet and (consequently) server domains Linux (and by extention the
rest of the open source community, or for that matter the whole rest of
the non-Microsoft-serf computer industry) is going to have to start
playing the same sort of follow-the-leader game that Linux has played on
the desktop?

The other (related) question I have is: why C#?  Again, I've never
seriously looked into C# other than the most publicly available of public
announcements, but so far as I can tell, C# is Microsoft's answer to Java.
It is supposed to be able to do what Java can do, only Microsoft gets to
control it and doesn't have to kowtow to anybody else.  Does C# have any
technical benefits in the specification of the language that Java does not
have?  Or could Java do everything that needs to be done, and is C#'s
*only* purpose to allow Microsoft to have something it made itself?

And, finally, the real question:  why are the Free Software solutions
going to be supporting C#?  Maybe there is a good reason: I hope to know
it.  But I suspect that this is going to turn out to be foolish.  Even if
C# specifications have been submitted to standards bodies by Microsoft, do
we *really* believe that the Microsoft implementation is going to continue
fully conform to these "open" specifications?  And if not, of course it's
going to be the Microsoft implementation's funcationality that forms the
"killer ap" for .NET.  Given all of this, I can't help but wonder if it
might be better for the Free Software community to build something with
all the functionaltiy and supposed benefits of whatever .NET is going to
be, but build it on top of Java, Python, or some other language, rather
than chasing after Microsoft's only pet standard.  That way, the community
will have the power to build the best possible system on tools which it
understands and has full access to.  Otherwise, the Free Software
community will forever be commiting itself to playing catchup with
Microsoft, and reverse engineering Microsoft's latest incompatable change
to some only-in-name "open" standard.

-Rob

From:	 Paul Winkler <slinkp23@yahoo.com>
To:	 letters@lwn.net
Subject: Mailing list for linux audio developers
Date:	 Sat, 07 Jul 2001 14:08:52 -0400

In the July 4 edition of LWN, I read this with interest:

"Developers of other languages should also consider the idea of this type of
cross-pollination effort. The idea could also be tried with other open-source
projects that involve parallel work on similar  projects. Areas that seem likely
to gain from such a collaboration include ... audio editor packages ..."

There is an existing mailing list that serves precisely this purpose for
developers of audio editors and other audio applications. The linux-audio-dev
list was started in 1998, and became quite lively in mid-1999. For subscription
info, archives, and other useful resources, see: http://www.linuxaudiodev.org

Current contributors include developers of some of most important (IMHO) audio
apps and frameworks on Linux today: alsa, Snd, Ecasound, GLAME, aRts, LADSPA,
Ardour, SoundTracker, sfront, csound, Sweep, GNU Octal, MusE, and probably many
others I've forgotten.

I think it is safe to say that linux-audio-dev participants have found the list
very helpful. It was on this list that LADSPA (the Linux Audio Development
Simple Plugin API) was conceived, debated, created, and finally became the de
facto standard for writing re-usable DSP code on Linux. LADSPA Previously, every
application either used its own plugin API or (more likely) none at all.  We are
now starting to see LADSPA support added to a number of existing apps that
formerly provided their own incompatible plugin APIs.

The hot topic at the moment is inter-application cooperation in a low-latency
realtime context. An API tentatively called LAAGA (Linux Audio Application Glue
Architecture) has been proposed, debated at great length, and now (thanks to
Paul Davis) is being tested in a reference implementation.

I suspect this message won't get posted on lwn.net until too late, but anyone at
LinuxTag in Stuttgart should be sure to stop by and check out the
linux-audio-dev demo booth! A number of list regulars will be on-hand all
weekend. Look for a blue-and-yellow logo with the initials LAD... or just follow
the interesting noises.


-- 
...................    paul winkler   ....................
custom calendars & printing: http://www.calendargalaxy.com
       A member of ARMS:   http://www.reacharms.com
            home page:  http://www.slinkp.com

From:	 "K.Hayen" <K.Hayen@digitec.de>
To:	 letters@lwn.net
Subject: ipfilter license
Date:	 Wed, 4 Jul 2001 15:04:57 +0200

I'd have expected more coverage on the issue of the license change. What is
attempted seems to be an attempt to protect a BSD project from GPL forks.
 
This is a valid interest. The license as is seems pointless, since GPL is
only a name and the new license doesn't address what GPL does.
 
I perfectly understand that's a unwishful situation to lead a BSD project
and see a fork happen that is still Free Software, but doesn't allow you to
merge back code into your own project, although the fork is just as well
publicly developed and thereby a direct competitor.
 
Maybe we will see a another license that is Free Software, but will forbid
sharing under other terms, but the ones given?
 
The GPL did that. And i love it for that. I don't see much of a problem in
BSD licenses, as well, FreeBSD ought to be GNU/FreeBSD, or not? And I don't
think it will gain more momentum than the GPL.
 
General purpose software will be almost 100% GPL some day. Just because it
makes sense money-wise for the users.
 
Yours, Kay

From:	 Rob Landley <landley@webofficenow.com>
To:	 letters@lwn.net
Subject: New IPFilter license.
Date:	 Fri, 6 Jul 2001 09:14:08 -0400

Your "security" section blurb on the new ipfilter license says:

>It resembles the BSD license, with one exception: it explicitly disallows 
>placing the code under the GPL. 

What it doesn't mention is that this "one exception" is something you can 
drive a mac truck through.  Here's the text of the "exception".

>The contents of this package may not be placed under the GPL or any
>other licence which requires requires [sic] you to give up your rights.

Any other license includes any proprietary license, which makes you give up 
the right to redistribute and modify which the license mentions earlier, so 
this code cannot be included in proprietary code.  In theory, any license 
that places additional restrictions on what you can do would be taboo, and 
since additional obligations are effectively restrictions (you can't do 
thing-one unless you do thing-two first), this basically prevents it from 
being relicensed at all except in the most superficial and cosmetic way.

It's interesting that the author of the license is incensed about the GPL's 
"restrictive" clauses preventing anyone from taking away the rights it 
guarantees, yet in objecting to it he created a license that does exactly the 
same thing on a practical level.  (Except that his license allows the 
distribution of binary-only versions, so the freedoms he aims to protect are 
not guarded in any practical way.)

In fact, his license is in a very real sense MORE restrictive than the GPL, 
because it bars code it covers from being integrated with a far greater 
amount of existing code (explicitly the existing installed base of GPL code, 
and implicitly anything under the MPL, artistic license, IBM's open source 
licenses, and countless others).

I'm not suprised he hates legalese.  He's not very good at it.

Rob

From:	 "Jay R. Ashworth" <jra@baylink.com>
To:	 letters@lwn.net
Subject: KDE v. Gnome
Date:	 Wed, 11 Jul 2001 01:17:34 -0400

[ Just a quick note, by comparison to some of my long diatribes :-) ]

I see that there's another wave of the "Gnome's better, KDE will
eventually die" nonsense coming in to shore this week.

I feel the need to remind all involved of one very important fact, which
seems to repeat itself all throughout history... and no one ever gets it:

	If you don't have an 'enemy', you don't get nearly as much
	accomplished, usually by at least one order of magnitude, if
	not two or more.

This hits hardest on US Military managers trying to get funded for mere
operations -- much less R&D -- in the wake of the "winning" of the Cold War,
but more productive advancement happened to SCO Unix since Linux hit the
radar screen than ever had before then, and the list of IT pertinent
examples is longer than I promised I'd write.  :-)

So, even if you think Gnome is the cat's ass, please remember that while
KDE may be merely toilet paper, the stuff *does* have its uses.  Ok?

Cheers,
- jra
-- 
Jay R. Ashworth                                                jra@baylink.com
Member of the Technical Staff     Baylink
The Suncoast Freenet         The Things I Think
Tampa Bay, Florida        http://baylink.pitas.com             +1 727 804 5015

   OS X: Because making Unix user-friendly was easier than debugging Windows