See also: last week's Back page page.

Linux links of the week

StepByStep is a different approach to providing Linux help and documentation. The StepByStep guides do not attempt to provide any sort of comprehensive coverage of a topic; instead, they are intended to be concise, quick guides to making something work.

Section Editor: Jon Corbet

Letters to the editor

From: Larry McVoy <lm@bitmover.com>
Date: Mon, 3 Jan 2000 18:08:22 -0800
To: editor@lwn.net
Subject: you might want to read this
Cc: lm@bitmover.com

[hold]

It appears to directly contradict what you are saying in

    http://lwn.net/2000/features/ncm-dvd.phtml

The following court case,

    http://www.law.emory.edu/7circuit/june96/96-1139.html

upholds shrinkwrap licenses, overturning a lower court's claim that
shrink wrapis not enforcable.

The basic summary is that the vendor can't do stuff like put a license
inside that says "because you opened the box, you now owe us another
$10,000, and paying us now is your only choice.  Ha ha, gotcha.".
However, the vendor _can_ put in the box, "your right to use this software
is conditional on you obeying the following rules (spell out the rules).
You can either agree to these rules or return your software for a full
refund."

In other words, a vendor can list rules, the court showed multiple
examples - from insurance policies to prescription drugs to software -
where such rules are listed and are expected to be obeyed.  In addition,
the court found that shrinkwrap does _not_ violate the UCC, as stated on
your web site.  The lawyer that OKed that web page appears to be sadly
misinformed about the state of the law.  And this isn't a recent case,
this is from '96.

--lm

To: letters@lwn.net
From: ncm@nospam.cantrip.org
Subject: Shrinkwrap Licensing

This is an update to my feature on shrinkwrap licensing,

  http://lwn.net/2000/features/ncm-dvd.phtml

in response to the LWN editors' and Larry McVoy's comments. 
LWN introduced the feature with a statement:

  Should the court rule on the validity of these licenses, it will be
  interesting to consider how free software licenses differ legally-- 
  if at all--from the commercial shrink-wrap variety.

Free Software licenses are based firmly on international copyright law.
The UCC (Uniform Commercial code) doesn't apply, because the copyright 
holders aren't selling you anything.  Red Hat doesn't own the copyright 
on (most of) the code in their box.  The UCC places obligations on 
Red Hat, but not anybody who is not party to the transaction, so the 
UCC doesn't weaken the GPL.

Larry McVoy introduces a more troublesome issue: the U.S. 7th Circuit 
Court overturned a district decision and upheld a shrink-wrap license:

    http://www.law.emory.edu/7circuit/june96/96-1139.html

The decision is troublesome because its reasoning is very sloppy, reading 
more like an undergraduate business-school essay than a serious legal 
document.  It dismisses the difference between a license and a contract  
in one line.  It similarly dismisses the very real practical problems of 
actually getting a refund after a product box is opened.  The examples 
the court takes as valid shrink-wrap licenses are drawn not from legal 
cases, but from other recent attempts at the same trick which happen not 
(yet) to have been fought all the way to a court decision.  The judges
note there is little case law, taking it to indicate that the public
implicitly accepts shrink-wrap licenses, despite that (as noted earlier)
software companies have routinely avoided trying to enforce such licenses
for fear of producing such case law.

Its basic argument is expediency: because it would be inconvenient for 
vendors to obtain agreement from customers to give up their rights under 
the law, it is sufficient (according to that court) for the vendor simply 
to assert that customers don't have those rights:  Not trying to return 
the product for a refund constitutes "agreement".  A customer who prefers 
to retain those rights has no recourse other than to try to get a refund 
(and good luck!).  The decision doesn't go so far as to say that a failed 
good-faith attempt at a refund might negate such an "agreement".

Fortunately for the DVD case, the 7th Circuit decision is (I believe) 
not binding in the 9th Circuit, where the DVD case is being tried.  
Furthermore, Norwegian law, which has jurisdiction where the 
reverse-engineering is said to have occurred, does not (according to 
Otto Skrove Bagge) allow a license to eliminate reverse-engineering 
rights.  Even if a contract-o-matic is held to constitute a valid 
contract, legally-invalid parts of such a contract are not binding.  
(Similarly, paragraphs common in real-estate title deeds in Los 
Angeles, forbidding sale to non-Causasions, are legally meaningless.)

The 7th Circuit precedent cries out for well-reasoned contradiction.
The DVD case might be an opportunity to evoke one, if only in passing.

I am not a lawyer, and the above has not been reviewed for legal accuracy.

Date: Tue, 4 Jan 2000 20:14:13 +0000
From: ruth@innocent.com
To: lwn@lwn.net
Subject: GPL as shrinkwrap license?


[HOLD]

As I understand it, the significant difference between say, the GPL and a
typical EULA shrink-wrap license from a major proprietary vendor is that
the GNU GPL is *not* an end-user license at all.

Only distributors and software developers need to agree to the GNU GPL,
the license itself says, in paragraph zero, "The act of running the
Program is not restricted (...)" because the architects of the GNU GPL
explicitly wanted everyone to be able to USE their software.

This means that for the purposes of the UCC, GNU GPL software does not
have any licensing restrictions applied to it. Purchasers of Gimp CDs
are free to use them as frisbees, install and use them on as many
machines as they like, and then re-sell the CD without any restriction.

Similarly, purchasers of a book are free to read it, use it to prop
up a table, discuss the plot with friends, then lend it to those same
friends and finally sell it second hand. Other rights are reserved to
the publisher, and there are extensive license agreements in place,
but like the GNU GPL they DO NOT MATTER to end-users.

Date: Thu, 30 Dec 1999 02:21:59 -0800 (PST)
From: Juergen Weber <weberjn@no-spam.yahoo.com>
Subject: Buffer overflow protection
To: letters@lwn.net

Hello,

in the security section of Dec,30,99
you write:

> But Linus's main point has always been that a
> non-executable stack is a band-aid solution which 
> does not fix the real
> problem - poorly written applications. 

In an ideal world where there are only wizards
like Linus you could fix the poorly written
applications problem.

The great inventions of computer science
made programming more error-prove.

Of course memory leaks are signs of poorly written
applications, but humans will always make mistakes.
So the java approach of freeing the programmer of
memory allocation was the way to go.

So the real solution is to disallow the execution code

on the stack.

Juergen


(please make my email address "anti-spammed")

Date: Thu, 30 Dec 1999 07:16:35 -0700
From: Ray Whitmer <ray@xmission.com>
To: letters@lwn.net
Subject: GNU/Linux

I read your recent item on the name of:  GNU/Linux versus Linux.  At
first, it sounded to me like a silly dispute.  But after reviewing the
GNU's page on this topic, I find that GNU's claims warrant
consideration.  Most developers understand that GNU has for years
supplied many pieces, and Linux was "only" a plugged-in kernel,
predating the Linux kernel by many years.  I believe GNU claims 28% of
the current size versus 4% in the Linux kernel, although it is not clear
to me which pieces they count.  There are also many other large valuable
parts of the combined O/S without which the kernel would be much less
useful -- I think especially of the XFree project.  It is not easy to
make sure everyone receives due credit as things evolve over time.
Calling it just Linux may seem to trivialize those other efforts.  While
it is not clear how to make this fair to everyone, perhaps in the future
multiple kernels will become available to plug in, and it will become
even more obvious than ever that the Linux kernel itself, while quite
important today, is only a small part.  I have friends who value FreeBSD
and other OS's -- there could be value in joining compatible parts of
various movements under a more-generic banner.

Ray Whitmer
ray@xmission.com

Date: Mon, 3 Jan 2000 12:34:02 -0600 (CST)
From: Dave Finton <surazal@nerp.net>
To: info@auroradev.com, stevef@auroradev.com, letters@lwn.net
Subject: Comment on your white paper


I do disagree with your assertions about Linux vs. NT.  In fact I will
highlight a few "innacuracies" in your white paper
(http://www.auroradev.com/whitepapers/open_source.htm) that need
to be addressed.  These aren't minor hiccups, but serious drawbacks to
what I was hoping would be a serious commentary on open source vs. NT

You said:

Since Linux is a network operating system, Linux may be a threat to
high-end NT, but not to desktop Windows: Microsoft Access, Office, and VB
development will continue to flourish independently of the network
architecture.

I reply:

NT *will* continue to flourish... even in the high-end and well as the low
end.  But you've seemed to miss the point that Linux is already
flourishing at all these levels as well.  Over 30% of all web servers use
Linux.  Linux is being taken seriously in our university (U of
Minnesota Duluth) ITSS department, which has always been staunchly
Netware-, Solaris-, and NT-centric.  Linux is around you in all
levels; you simply fail to see it.

Also Linux is booming on the desktop.  People who I didn't even think
would consider using it (i.e. the "average joe") are telling their friends
they've installed Red Hat or Mandrake Linux on their machines and really
like it.  And guess what?  They're *using* it too, in increasing
numbers.  Frankly, your hypothetical situation does not exist in the real
world.

You said:

The Palm Pilot was the last great hope in the anti-Microsoft
camp. Microsoft responded with Windows CE, a lean and mean operating
system designed to run on handheld computers, palm devices, car radios,
and cell phones. This great new OS supports color screens (where is that
color Palm Pilot?) and much superior handwriting recognition. There are CE
versions of all your favorite Office Products, and a Visual Basic
developer?s kit for CE. I could not tell you how to wirte an application
for Palm Pilots (not even Java with its Write once Run many fame can run
on a Palm without major modifications to the core language.), however, I
can create a Pocket Access or VB application for the CE in minutes.

I reply:

Ah, so that's why Palm Pilot still is beating CE in virtually every market
I know of, and has been doing so for *years*.

Portability of apps to the CE devices cannot overcome Windows' flaws on
the handheld devices.  Palm is simply better.

You said:

Most Linux installations in production are UNIX shops that run $160,000 +
SUN servers. You can get the same power, scalability and performance with
Microsoft Cluster Services and NT for one third the price!

I reply:

Uh, Linux is free, and it comes with the same power, scalability, and
performance with Beowulf Cluster services and Linux comes at zero thirds
the price!

Linux is making *serious* inroads into many markets, particularly in
formerly NT-centric shops where Microsoft has disappointed IT managers one
too many times.  NT is simply not a cure-all.  Mind you, neither is Linux,
but I can't agree with your premises here.

I find your "white paper" technically misleading and innacurate, and
cannot take it seriously as such.  I have to implore you to take a serious
look into the marketplace.  The results may surprise you.

                          - Dave Finton

p.s.  On a final note, you're probably assuming that Microsoft is
unbeatable.  Remember when they said the same about IBM and DEC?  Nobody's
glory years last forever.

---------------------------------------------------------
| If an infinite number of monkeys typed randomly at    |
|   an infinite number of typewriters for an infinite   |
|   amount of time, they would eventually type out      |
|   this sentencdfjg sd84wUUlksaWQE~kd ::.              |
| ----------------------------------------------------- |
|      Name:      Dave Finton                           |
|      E-mail:    surazal@nerp.net                      |
|      Web Page:  http://surazal.nerp.net/              |
---------------------------------------------------------

Date: Thu, 30 Dec 1999 08:17:56 -0500
To: letters@lwn.net
From: "Gregor N. Purdy" <gregor@focusresearch.com>
Subject: Fwd: An idea

LWN--

I sent the following to the FSF after reading about the Amazon.com
boycott.

---------------------------------------------------------------------

Someone else has to have thought of this, but I haven't run across
it anywhere in my "travels" yet.

As long as the current PTO stance remains uncorrected, things like
this will happen. If the past is any guide to the future, any
correction to this will take a long time. Therefore, while pursuing
a correction to the policies and practices of the PTO is vital, we
should be looking for ways to relieve some pain in the interim.

I suggest that we seek out "Angels" in companies that are making
money from free software, such as Red Hat and VA Linux, and via
fundraising through LPF and GNU to fund a legal entity that will
file for and defend patents with automatic free license granting
similar to the provisions of the GPL.

So, we can put together patent applications for important techniques
that we fear will be stolen from the community by companies through
inappropriate PTO usage. Once patents are granted, usage of the
techniques will fall under the license agreement mentioned above,
which will state that no entity holding software patents that are
not licensed under this license may use the technique (not even for
a fee).

As the portfolio builds, and hopefully with a lot of help from the
commercial folks who live by free software both in funding and in
generating patents, we can start to carve out some free territory.
And, by setting an example, hopefully we can pull in other companies
that *want* to play with the free software folks, convincing them
to either (a) transfer their patents to this other entity or (b)
retain official ownership, but permanently license them according
to the GNU Intellectual Property License (GIPL), or whatever the
thing is called.

Stop software patents! But, in the mean time, take some defensive
action. Besides, this would probably bring a lot of attention to
the issue...

If companies are creating the patents in order to get recognition
of their achievements (an idea which fails for the "simple and obvious"
category), then there shouldn't be any issue subsequently licensing
them this way or transferring ownership. The free software community
has always been big on giving credit where credit is due (and only
where due). For those companies doing it for the purpose of protectionism
(weak) or extortion (evil), hopefully we can create an uncomfortable
environment for them between now and when the problem is fixed.

--Gregor N. Purdy
Focus Research, Inc.
gregor@focusresearch.com

Date: Thu, 30 Dec 1999 12:29:07 -0500
From: atorrey 
To: letters@lwn.net
Subject: Thoughts on the Amazon boycott

 This is likely to get me flamed, but I have serious doubts about the
potential effectiveness of the Open Source community's call for a
boycott of Amazon.com over the One Click software patent. 

 While we like to tell each other how 'special' we all are, the
practical hard fact is that the Open Source world, even if we include
all our friends, is not all that big a percentage of the world.  Even
the most successful boycott is unlikely to have a major impact on
Amazon's bottom line.  (A similiar logic could most likely be applied to
other calls for boycotts of other companies for S/W patents)  Indeed,
Amazon is big enough that there are often few on-line alternatives,
especially if one also finds Barney igNoble obnoxious.

 While I am not saying to dump the boycott, to me it is not using our
talents to their best advantage.  Amazon is a marketting specialist, and
a boycott is trying to beat them at marketing, why tackle them on their
own turf?  If you want to beat someone, it is best to work from your
strongest position.

 Why not let Amazon keep their patent, just like we let another major
corporation keep it's O/S, and go for 'world domination' with our
strongest skill set.  Go to Barnes & Noble, and offer to help them
develop a non-patent infringing, improved, equivalent to One Click. 
(One possible idea - how about if the system enabled a single checkout
from an entire shopping session, involving multiple e-stores?)
Obviously it would have the string attached that it would be Open
Source, perhaps with a limitation in the liscence that it was only open
to companies that did not use software patents to limit competition...

 Which idea would you think would make Jeff Bezos more nervous - the
thought that a few hundred geeks might take their business elsewhere? 
Or that those same geeks, who include some of the worlds best
programmers, are going to go help the competion build a better website?

 ART

  (Please do not include my E-mail address, if you must, please
anti-spam it...)