LinuxSA Mailing list archives

  From: Alan Kennington <akenning@topology.org>
  To  : Martin Sandiford <ms@mcdev.com.au>
  Date: Sun, 6 May 2001 14:56:59 +0930

Re: request for advice on non-free open license

With permission, I'm copying some of Martin Sandiford's
very helpful response back to this list.

===========================================================
On Sun, May 06, 2001 at 11:36:18AM +0930, Martin Sandiford wrote:
> On Sun, 6 May 2001, akenning@topology.org wrote:
> > Can I impose commerical licence conditions on 2 source 
> > files which are to be added to a linux kernel to 
> > extend its functionality?
> 
> Probably not, unless it can be loaded/unloaded as a module using
> insmod/rmmod.
> 
> > Question:
> > Is it in accordance with the letter of the GPL to add my own
> > completely original code to the kernel - as extensions of linux -
> > and insert a copyright notice and licence notice saying that
> > the user must pay me money?
> 
> I don't think that this is either within the letter or the spirit of the
> GPL.  In my opinion, this is *exactly* what the GPL is trying to
> prevent.

Martin,

Fair enough.
My instinct is always to publish first, and think about my loss of
income second. Since I spent so much time in academia, I agree with
Linus that Isaac Newton was an open-sourcer who used from the open source
and gave back to it. On the other hand, he was a genius, and he had
uni tenure, and patents didn't exist then, and a theory of gravity
wasn't likely to be commercialisable at that time.

I still think that the right way to develop software is to start up
"software universities" worldwide, which are funded by governments
to create free open source software. Any govt which did this would
greatly increase their pool of software skills, and the economic
benefits would be substantial. In the current climate, Isaac Newton
would have been required to patent his theories, hand over ownership
of the patents to the Uni, and help set up a joint venture with
industry to exploit it. And the uni administrators would have
investigated whether there was a sufficient market for graduates
who are experts in theories of gravity and motion.
But I digress...

> > If it was a patch, I imagine that it _would_ be forbidden by
> > the GPL to ask money for the patch.
> 
> No.  The GPL does not prevent anyone from asking money for code.  It
> does prevent anyone from not sharing the code.  It also does prevent
> things like "you can't have or use this code if you don't pay me the
> money".

In my case, I'm thinking of linux as a splendid publication medium.
Once something goes into the linux kernel, it's prior art.
And that pre-empts those people who think that knowledge should
be private property.

> > In my case, it's sort of a module, but not a device driver style
> > of module. I.e. it's a module in the engineering sense, but not
> > in the sense of having a major device number sort of thing.
> 
> For Linux purposes a module:
> 
> 1. Can be compiled independently of kernel sources
> 2. Can be loaded and unloaded at run-time (using insmod/rmmod)
> 
> If it isn't this kind of module, *and* it forms part of the kernel, then
> it needs to be under GPL.

That's crystal clear.

> There are a couple of things that you could do to release it under a
> non-GPL license that suited your purposes better:
> 
> 
> 1. Write it as a real Linux kernel module.
> 
> If the current kernel does not provide the module interfaces required to
> support your code as a Linux module, then you could implement the kernel
> interfaces as a kernel patch, and release this patch under GPL.  You
> would then be free to license your code (which is now run-time loadable)
> under whatever license you choose.

Now that's a really sneeky idea.
I'll keep that option in mind.

> 2. Target an O/S that does not require you to release patches and/or
>    drivers as GPL.
> 
> These operating systems include FreeBSD, NetBSD, OpenBSD.  If your code
> provides functionality in an internet server context, then the *BSD
> platforms provide an arguably more secure and stable platform (my
> flame-proof undies are on :) ).

I think that the BSDs don't have support for the things I want to do.
But I'll look into it.

> Microsoft Windows variants might also be in this category as far as
> licensing issues are concerned.  I'm not convinced that Windows provides
> a stable or secure internet platform though.

Blurk!

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

There are another couple of options I thought of.

1.	Write obfuscated code, more obfuscated even than
	sendmail.cf language (or algol).
	But I remember that the GPL forbids that sort of thing.

	On the other hand, linus actually says in his style guidelines
	that authors should avoid comments.
	As a result, when I had to debug a device driver a year ago,
	I had to wade through 3000 lines and comment it all
	myself to make sense of it.
	Non-commenting is fine if you never intend anyone else to
	read it, and you don't intend to come back to it a year later.
	But I find that intensive internal documentation is essential.

	My point is that effectively, linus actually encourages
	code obfuscation. He does say that the code should be
	self-explanatory. But what he calls self-explanatory,
	I call baffling and inscrutable.

2.	I bundle up the software in a box, embedded in EPROM
	or something like that, and no one gets to see it.
	Then I'm not selling the software, I'm selling the box.
	But that might not be permitted either.

3.	I publish the source on a website which no one ever visits.
	And I don't tell anyone what it's useful for.

I think the real problem with the open source software is that it is:

-	fine for users who get free+open source.
-	fine for students who write for nothing, for hobbyist
	programmers, and for big companies which make their money
	from something else.
-	bad for people who want to make software development
	their primary activity, with nothing else to
	subsidize this activity.

No one has yet worked out how you can get a fair day's pay for
a fair day's open source development.
(All the stuff I've given away so far just results in requests
every week for free assistance!)

----------------------------------------------------
Anyway, thanks very much for clarifying things.
I think I'll "published and be damned", as they say.

Cheers,
Alan Kennington.

-- 
LinuxSA WWW: http://www.linuxsa.org.au/  IRC: #linuxsa on irc.linux.org.au
To unsubscribe from the LinuxSA list:
  mail linuxsa-request@linuxsa.org.au with "unsubscribe" as the subject