On Saturday, August 16, 2003, at 11:09  AM, Joshua 'Schwa' Gargus wrote:

> On Sat, Aug 16, 2003 at 08:11:56PM +0300, Daniel Vainsencher wrote:
>> As Andrew once said, absolute legal safety doesn't exist. As it 
>> appears,
>> software patents are soon going to be a general problem for the 
>> software
>> world to deal with. We are even less likely to find a complete 
>> solution
>> to false accusations or honest slip ups.
>
> Fair enough, but that only addresses one of Cees' points.  It would be
> very difficult for anyone in the Squeak community who wrote (say) a
> replacement VM to argue that they had done it in a "clean room".
> Therefore, a rewrite doesn't really buy us much.  Right?

Three points here:

First, a rewrite buys us a better VM. I can't imagine anyone would 
write a new VM just to change the license. I also can't imagine anyone 
would do that much work and not take the opportunity to make big 
improvements.

Second, I don't think there *is* such a thing as a perfect 'clean room' 
in Smalltalk. You can do it with say, BIOS. But anyone who knows enough 
to implement a VM or a kernel is undoubtedly familiar with somebody 
else's implementation. Who's to say that the similarity isn't with, 
say, VW or ST80, rather than Squeak? For that matter, how could anyone 
claim that the original work at Apple wasn't influenced by anything 
else?

Third, I think the notion that "absolute legal safety doesn't exist" 
applies even more to rewrites than to patents. We cannot achieve 
immunity from lawsuits. The question is how much legal risk are we 
willing to take? Note that right now, we running some level of legal 
risk that isn't terribly well identified.

Colin