Re: [oc] Re: CAN core

[Richard Herveille <richard@asics.ws>]

On Wednesday 21 August 2002 04:08, Andreas Bombe wrote:
> On Tue, Aug 20, 2002 at 03:50:58PM +0200, Richard Herveille wrote:
> > > I'd say that's better
> > >
> > > I'l write to them on this issue
> > >
> > > So what I make out is that it would not matter if we make the core
> > > but when some one uses it for commercial use they have to pay
> > > Bosch for that
> >
> > Yes. If you write a core (from scratch) based on freely available
> > information (like the I2C datasheet), you should be ok. But if you start
> > reverse-engineering something, you might get into trouble.
>
> If there is a patent on some part of CAN (and it sounds like there is),
> then it doesn't matter whether you reverse engineer something or build
> it from publically available documentation.  Even when you've never
> heard of CAN before and reinvent it, the patent still applies.
>
> After all, it's the whole point of a patent to encourage publication of
> technological methods (in exchange for protection of the invention).
> Designing to spec doesn't circumvent it.

Correct, what I tried to say is that it usually is allowed to create cores for 
patented designs. But that the patent still holds; somebody has to pay the 
fees (usually the chip manufacturer).

The whole idea of a patent is to provide technology inventions to the public, 
and to let others improve the technology. The basics are patented by the 
first inventor, the improvements can/may be patented by the second inventor.

MP3 is another great example. The MP3 system is patented it is owned by the 
Fraunhofer Institut fur Integrierte Schaltungen. But look at the amount of 
MP3 devices out there. Fraunhofer's doesn't want to keep MP3 a secret. Hell 
no, they want everybody to know how it works, and they want everybody to use 
it. The more people using it, the more fees they can collect.

The same is true for Philips and the I2C bus. And my guess, the same is true 
for Bosch and CAN. As long as the user pays the fees, everybody is happy.

Richard
--
To unsubscribe from cores mailing list please visit http://www.opencores.org/mailinglists.shtml