Re: Response from M-Systems

Re: Response from M-Systems

[David Woodhouse]

jgg@deltatee.com said:
>  Ah, then it gets complex, since you don't have any patent right to
> give you have contradicted the terms of your own license. The
> conservative view is that anyone you give it to has no license to use
> it because they cannot comply with the license, 

I honestly can't see how. Much as I disagree with the concept of software 
patents, and I hope we don't have them inflicted upon us here in the UK - I 
can't see what the GPL's restriction is. I'm aware of the _intention_, but the 
GPL doesn't actually appear to assert the the code must be _usable_ by anyone. 

In fact, it explicitly states that it's only about "copying, distribution and
modification", and the patent doesn't attempt to restrict you from either of
those three activities; only from _using_ the code with non-M-Systems'
hardware. You can do what you like with it, but as soon as you actually start 
using it on hardware that's not produced by M-Systems, you're violating the 
patent.

Yes, this is a contentious statement for me to make, and one which runs 
contrary to the spirit of the GPL, but I really can't see where the flaw is in 
this logic.

Also, consider the context. I'm talking about releasing code myself under GPL 
- not suggesting that M-Systems do so. I'm trying to convince them that it's 
safe to allow me to do so; and to assist me by giving me specs. So this 'flaw' 
in the GPL actually works to our benefit in this case.

If someone duplicates their hardware, and it comes to court, then there are 
two possible scenarios:
 1. My reasoning is shown to be correct - M-Systems win the case.
 2. My reasoning is shown to be incorrect. Results:
	I was never allowed to release the code under GPL in the first place.
	--> The GPL license on the code the competitor is using is invalid.
	--> The competitor is using unlicensed code.
	--> The competitor is _also_ violating M-Systems' patents.

So either way, M-Systems have no reason to avoid giving me the specs for the 
hardware and allowing me to release GPL'd code. It's good for us and it's good 
for them. Worst case: The GPL on the code gets retracted at some later date, 
after we're all using it.

There's also the fact that any reverse-engineering would lead to BSD-licensed 
code, just to spite them - so releasing GPL'd code is in fact a way for them 
to _restrict_ the terms of the source distribution in the long term.


----                                 ----                                 ----
David Woodhouse        David.Woodhouse@mvhi.com       Office: (+44) 1223 810302
 Project Leader,     Process Information Systems      Mobile: (+44) 976 658355
    Axiom (Cambridge) Ltd., Swaffham Bulbeck, Cambridge, CB5 0NA, UK.
             finger dwmw2@ferret.lmh.ox.ac.uk for PGP key.




To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[Richard Stallman]

A couple of months ago, I asked our lawyer a similar question: whether
someone could permit distribution of a GPL-covered program, but then
use a patent to try to restrict the running of the program.  Such a
possibility would be disastrous, since it would mean that the program
was not really free, despite the use of the GPL.

The lawyer told me that licensing the distribution of a program
automatically licenses users to run it.  I was relieved.

I cannot be certain that the case I asked him about is entirely
equivalent to this one.  What I can say is that IF a program can be
restricted in the way you suggest, it is not free software.  That is
not a solution, it is a danger.

I don't think this danger exists, but if it does, I will do my best
to safeguard against it with changes in the next version of the GPL.


To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[David Woodhouse]

rms@gnu.org said:
>  The lawyer told me that licensing the distribution of a program
> automatically licenses users to run it.  I was relieved.

That seems very strange to me - I'd be very interested to know why he believes 
this. Distribution and usage are two distinct activities; one of which is 
explicitly allowed by the GPL, the other of which is explicitly not covered by 
the GPL. 

> I cannot be certain that the case I asked him about is entirely
> equivalent to this one.  What I can say is that IF a program can be
> restricted in the way you suggest, it is not free software.  That is
> not a solution, it is a danger. 

This is only with regard to software patents, a local issue in some parts of 
the world which have insane legislation.

If a tinpot dictatorship somewhere ruled that source code should never be 
available to the end user, the GPL would be unusable in that country, and by 
the same logic, no software anywhere would be truly free. But would we stop 
using the GPL?

What I'm trying to say is that software patents are a local issue. GPL'd code 
is truly free in all parts of the world with sensible legislation, and in 
those parts that aren't included in that description, we have to fight it in 
other ways. We can't just say 'well the software's not free' and reject it 
outright. 



----                                 ----                                 ----
David Woodhouse        David.Woodhouse@mvhi.com       Office: (+44) 1223 810302
 Project Leader,     Process Information Systems      Mobile: (+44) 976 658355
    Axiom (Cambridge) Ltd., Swaffham Bulbeck, Cambridge, CB5 0NA, UK.
             finger dwmw2@ferret.lmh.ox.ac.uk for PGP key.




To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[Richard Stallman]

    >  The lawyer told me that licensing the distribution of a program
    > automatically licenses users to run it.  I was relieved.

    That seems very strange to me - I'd be very interested to know why he believes 
    this.

Lots of things in the law seem strange to non-lawyers.
His answer was a (pleasant) surprise to me, too.

    What I'm trying to say is that software patents are a local issue. 

Local to the United States, the UK, Japan, and maybe a year from now
all of Europe.  It is rather callous to dismiss such a large part of
the computer-using world as "local".


We can offer a company like M-systems certain kinds of cooperation
if they cooperate with our community; we can offer this cooperation
in the context of our relationship with them specifically.
But we cannot make basic global decisions for the sake of
our relationship with one company.


To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[David Woodhouse]

jgg@deltatee.com said:
>  I'm not sure you can actually patent and enforce the software side of
> a hardware interface. Ie if we write a driver that does the proper
> sequence to page, write and erase their chip they cannot take action
> agaist us, unless perhaps that sequence involved lots more software
> control that the AMD flash chips do.

That I suppose I'd agree with - although I would like to get NFTL working too, 
which is a different matter.

> Can you dig up the patents they actually hold on their FTL and on
> their hardware interface? 

http://www.patents.ibm.com/ and search for "M-Systems". I can't see NFTL, just 
FTL, which is a moot point because it's already done. There's some stuff on 
hardware, too.

----                                 ----                                 ----
David Woodhouse        David.Woodhouse@mvhi.com       Office: (+44) 1223 810302
 Project Leader,     Process Information Systems      Mobile: (+44) 976 658355
    Axiom (Cambridge) Ltd., Swaffham Bulbeck, Cambridge, CB5 0NA, UK.
             finger dwmw2@ferret.lmh.ox.ac.uk for PGP key.




To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[David Woodhouse]

David.Woodhouse@mvhi.com said:
>  Worst case: The GPL on the code gets retracted at some later date,
> after we're all using it. 

I ought to have pointed out that this applies to _any_ GPL'd driver for the 
Disk-On-Chip and NFTL, regardless of whether it's produced by M-Systems, with 
M-Systems' specifications, or by reverse-engineering.

If we accept the premise that the patents prevent the code from being released 
under GPL, then what are we playing at in the first place? 

----                                 ----                                 ----
David Woodhouse        David.Woodhouse@mvhi.com       Office: (+44) 1223 810302
 Project Leader,     Process Information Systems      Mobile: (+44) 976 658355
    Axiom (Cambridge) Ltd., Swaffham Bulbeck, Cambridge, CB5 0NA, UK.
             finger dwmw2@ferret.lmh.ox.ac.uk for PGP key.




To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[Jason Gunthorpe]

On Fri, 9 Jul 1999, David Woodhouse wrote:

> I ought to have pointed out that this applies to _any_ GPL'd driver for the 
> Disk-On-Chip and NFTL, regardless of whether it's produced by M-Systems, with 
> M-Systems' specifications, or by reverse-engineering.
> 
> If we accept the premise that the patents prevent the code from being released 
> under GPL, then what are we playing at in the first place? 

I'm not sure you can actually patent and enforce the software side of a
hardware interface. Ie if we write a driver that does the proper sequence
to page, write and erase their chip they cannot take action agaist us,
unless perhaps that sequence involved lots more software control that the
AMD flash chips do.

Can you dig up the patents they actually hold on their FTL and on their
hardware interface? 

Jason



To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[David Woodhouse]

jgg@deltatee.com said:
>  Yeah, he wouldn't. The GPL is reasonably clear that you have to give
> up patent rights. IBM's laywers and Apple's laywers both concure on
> this. If you as a compay GPL an implementation for an algorithm then
> that company must give free use patent rights to everyone who recieves
> a copy of the code, which is effectively everyone. 

And if I personally release some code under GPL, what then? Have I 
singlehandedly voided their patents, to which I didn't have any right in the 
first place?

I'd like to know precisely why IBM and Apple lawyers think that. I can't see 
it myself. Of course, I'm not a lawyer, but I'd expect to understand if I saw 
the explanation. Can you see an actual flaw in my logic?



----                                 ----                                 ----
David Woodhouse        David.Woodhouse@mvhi.com       Office: (+44) 1223 810302
 Project Leader,     Process Information Systems      Mobile: (+44) 976 658355
    Axiom (Cambridge) Ltd., Swaffham Bulbeck, Cambridge, CB5 0NA, UK.
             finger dwmw2@ferret.lmh.ox.ac.uk for PGP key.




To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[Jason Gunthorpe]

On Fri, 9 Jul 1999, David Woodhouse wrote:

> And if I personally release some code under GPL, what then? Have I 
> singlehandedly voided their patents, to which I didn't have any right in the 
> first place?

Ah, then it gets complex, since you don't have any patent right to give
you have contradicted the terms of your own license. The conservative view
is that anyone you give it to has no license to use it because they cannot
comply with the license, a less conservative view is that they cannot
distribute source or binaries derived from it because they cannot comply
with the license. You can distribute it because you can't really break
your own license :>

Basically most patent rights are incompatible with the GPL so you end up
with the exact same mess as the KDE people have, code that is GPL'd and
code that is not GPL'd (via the patents) - sorting it out is not trivial.

Jason



To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Response from M-Systems

[David Woodhouse]

I've heard again from Amir Ban, the Vice-President of R&D at M-Systems. He
presents M-Systems concerns and reasoning for their behaviour. It remains to be
seen whether this is a final response or whether they're still trying to reach
a more useful conclusion.

My response to their concern about patent issues, of which RMS is probably 
not going to approve, is also on the page:

	http://www.linux-mtd.infradead.org/msys-response.html



----                                 ----                                 ----
David Woodhouse        David.Woodhouse@mvhi.com       Office: (+44) 1223 810302
 Project Leader,     Process Information Systems      Mobile: (+44) 976 658355
    Axiom (Cambridge) Ltd., Swaffham Bulbeck, Cambridge, CB5 0NA, UK.
             finger dwmw2@ferret.lmh.ox.ac.uk for PGP key.




To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk

Re: Response from M-Systems

[Jason Gunthorpe]

On Fri, 9 Jul 1999, David Woodhouse wrote:

> I've heard again from Amir Ban, the Vice-President of R&D at M-Systems. He
> presents M-Systems concerns and reasoning for their behaviour. It remains to be
> seen whether this is a final response or whether they're still trying to reach
> a more useful conclusion.
> 
> My response to their concern about patent issues, of which RMS is probably 
> not going to approve, is also on the page:
> 
> 	http://www.linux-mtd.infradead.org/msys-response.html

Yeah, he wouldn't. The GPL is reasonably clear that you have to give up
patent rights. IBM's laywers and Apple's laywers both concure on this. If
you as a compay GPL an implementation for an algorithm then that company
must give free use patent rights to everyone who recieves a copy of the
code, which is effectively everyone.

Given his statements I would make 3 assertions
  #1 - We are already cloning your higher level software technologies, if
       you with hold hardware specs then not only will your competitors
       have the software that you are afraid of giving them, but your 
       customers also won't have support for your hardware on our
       platform! You loose both ways.
  #2 - Your hardware patents protect you from people cloning the actual
       hardware device

I can state from experiance that we have looked at using disk on a chip
here and rejected it because of their attitude towards us needing custom
drivers for QNX.

Personally I don't the NFTL is very interesting, the specs to access the
r/w/e the DoC hardware is what is important and probably is pretty trivial
to reverse engineer, stick a logic analyser on the chip's bus and voila :>

Jason





To unsubscribe, send "unsubscribe mtd" to majordomo@imladris.demon.co.uk