Re: APSL 1.1 certification... [talking with the lawyers]

On Sat, Apr 24, 1999 at 11:24:07PM -0700, Brian Behlendorf wrote:
> Hi folks. The issues raised in the last week on this list about the APSL
> 1.1, in particular about the language used in section 9.1, have been very
> persuasive in our consideration of whether to grant OSI certification.
> Because of that, we have not arrived at a consensus on whether to certify
> yet, or deny it.

Hi,

I have had a few conversations with our firm about using code that is
released under the APSL and thought that the results of our
discussion might be somewhat useful. So, I'll try to sum up some of
the thoughts that we came up with and see if anybody thinks it makes
any sense.

The first thing that talking with our attornees made clear to me was
that Apple has a valid concern in the case of contributory infringement.
Dale (my attorney) told me that he was currently involved in a suit
in which a bag manufacturer had been drawn into a patent suit. In
this case, when this particular bag is empty it is not infringing on
anyone's patents but when you fill it with certain materials it is.
So, this bag company was being sued because they were selling bags to
someone that they knew would fill them with the kind of materials
that would constitute a patent violation. Thus, the bag maker was
sued. If this kind of thing holds water in a patent suit then it is
crystalline to me that Apple should feel "the ph34r" in a big way.

So, with that in mind, I can see why Apple wishes to add section 9.1
and the other associated language about patent infringement.
However, I think that there is some work that can be done to make
Apple's intentions more clear and hopefully make the license a little
more fair to the people that will derive work from it or improve it.
One of the clearest problems to me is the lack of any definition of
context in what does and does not constitute Affected Original Code.
As the license stands it isn't clear to me whether this is a global
phenomena or a local one. Let me try to elaboarte a little on that.

In an earlier message I talked about a fictional situation where
a piece of encryption code that Apple had released under the APSL
became the subject of an infringement suit. In this fictional
scenerio, the use of the encryption in the context of a web server
was considered a violation and the use of it in a messaging protocol
was not. Apple had produced a web server with the encryption and had
thus been drawn into a patent suit. Apple, therefore, exercises their
right to terminate the use of the effected code and I am prevented
from using the code in the context of my messaging protocol. Brian
pointed out that the part of 9.1 that allows me to make my own
arrangements ("nothing in this license shall be construed to restrict
you...") but didn't address the fact that Apple is the only party
that I need to negotiate with. After all, I'm not in violation of
any patents, I just don't have a right to use Apple's code anymore.

The more interesting part of our converation was when we explored a
reversal of this situation. Let us say instead that the patent suit
revolves around my use of the encryption code in a messaging protocol
and not Apple's use in the web server. Now, lets say that this
company that sues me for using the code figures out that I have no
money to speak of and decides to go after Apple for contributory
infringement just like they went after the bag maker in Dale's suit.
Good thing for Apple that they can terminate my license and maybe dodge
that bullet. But, the downside is, does my having infringed on that
patent now terminate everyone's rights to use the "Affected Code"
regardless of the context of their use? Under the current agreement
there is nothing that makes any notion of context of use an apparent
factor. Affected Original Code is just the segment of Original
Code that has been referenced in a patent suit and is not necessarily
limited to the use of that code in a specific infringing context.

It seems to me that the license is very close to being a good license
for the community but fails in just a few key areas. I think that if
we could introduce a concept of context to the definition of Affected
Original Code that it would go a long way towards reassuring the
potential users of this codebase. In other words, if a piece of APSL
code becomes the subject of patent litigation then the right to use
that code _within_that_violating_context_ is terminated. Other
programs which use the same Original Code but in a context that does
not constitute patent infringement would be unaffected.

Does anyone follow that?

--
___________________________________________________________________
Ean Schuessler Director of Strategic Weapons Systems
Novare International Inc. A Devices that Kill People company
*** WARNING: This signature may contain jokes.