On Mon, Dec 19, 2005 at 10:17:58AM +0100, Wouter van Ooijen wrote:
> > Darn, I was going off of their original plan which was to use GPL 
> > licenses or similar.  But now the site isn't responding from here and 
> > you have me curious if they were able to stick with it.
> 
> The problem with GPL + hardware is that GPL talks about things like
> combining into an application, which is more or less well-defined when
> talking about software. But how would you interpret that when deling
> with hardware? That is a very important question, because it is the
> boundary of the GPL viral/contaminating effect. With software the
> boundary is the linked application. But were would the boundary be with
> hardware? Chip level? PCB level? Apparatus level?

I'll take a stab at it.

First off GPL probably doesn't fit unless you are releasing a complete
solution. In that case then all modiciations to the complete solution
would need to be published. However if you're talking about a component
architecture, then an LGPL level license is more appropriate. However, as
I stated in my other post it has its issues. The LGPL has three basic
tenets:

1. Changes to the LGPL component itself is treated as GPL. As such 
modifications to LGPL components must be published.

2. Developer generated components are generally not subject to license
restriction. So the developer can license their component any way they 
like.  In addition the combined work can be licensed as the developer sees
fit. There is one exception...

3. Since the LGPL component is considered to be free, the end user
retains the right to update/improve the LGPL component even when that
component is embedded in a complete solution with other non LGPL
components. In addition developers who use the LGPL component must
provide facilities for the end user to update/impove the LGPL component.
This can be up to and including the publishing of the developers components
if they are required for this update/improvement to occur.

So take for an example a LGPL ASIC that's integrated into a larger board.
As an LGPL component, any changes made to the VHDL of the ASIC would be
need to be published. However, none of the other components would necessarily
have to be published. On the other hand, the developer still must provide
facilities (JTAG interface for example), so that the end user can update
the ASIC.

Personally I have a problem with the third tenet. While I understand that it
keeps the LGPL component free from a principle standpoint, there's just too
much baggage associated with it. A streamlined license that keeps the first
two tenets and ditches the third would be a good balance. End users would no
be able to update and developers would be able to keep their code to
themselves while having to publish changes to the LGPL component.

BAJ
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist