Patrick asked me to forward this. ---------- Forwarded message ---------- From: Patrick McNamara <[EMAIL PROTECTED]> Date: May 18, 2005 10:08 AM Subject: Re: [Open-graphics] Whitepaper on releasing the RTL as open-source. To: Timothy Miller <[EMAIL PROTECTED]>
--- Timothy Miller <[EMAIL PROTECTED]> wrote: > On 5/18/05, Lourens Veen <[EMAIL PROTECTED]> wrote: > > On Wednesday 18 May 2005 02:00, Timothy Miller wrote: > > > On 5/17/05, Eric Smith <[EMAIL PROTECTED]> wrote: > > > > Timothy wrote: > > > > > > > > I'm hoping that Timothy and his partners and investors will be willing > > > > to make the RTL available *before* it is GPL'd to hardware hackers that > > > > have a serious interest in contributing to the project, under a non-GPL > > > > license that doesn't allow public distribution. If there is a serious > > > > commitment to releasing under the GPL by a certain date (especially if > > > > there is escrow), that should provide sufficient access for people to > > > > start contributing. > > > > > > With an NDA, this should be fine. Of course, there's a risk of it > > > getting leaked if too many people get it, so even with the NDA, it > > > shouldn't be free; this way, only serious people get it, and I believe > > > there are some legal reasons why a contract is more binding when money > > > is exchanged (something about 'consideration'), but IANAL. > > > > > > In any event, what should it cost? And it would be easier to roll the > > > 'hobbyist' and 'commercial' license into one, where the up-front fee > > > isn't too bad in either case, and there's also a royalty for each chip > > > you produce. > > > > > > Suggestions? > > > > Keep them separate. Have a commercial licence at a commercial rate (I think > > $25k was mentioned, I have no idea what would be reasonable), and have a > > hobbyist licence that is much cheaper, but requires the licencee to > > contribute the source to his/her derived works back to the project if they > > are distributed (like the Mozilla Public License), that is if you allow > > distribution in binary form of derived works at all. This way, hobbyists > can > > hack privately, and contribute to the project, but if you want to use the > IP > > in your own proprietary project you'll need to get a developer's licence. > > That sounds good. Anyone else have any objections? > Tiered licensing is a good idea. Here is a set of suggested limitations of the license types. This isn't a complete list, nor is it anything other than me thinking of the top of my head. Commerical: FPGA and ASIC RTL Binary redistribution allowed. Source registribution allowed. Re-licensing allow with same of more restrictive terms Design level improvement must be contributed back RTL level improvment may be kept proprietary. Very high license cost. Educational Institution: FPGA and ASIC RTL Binary redistribution of ASIC RTL not allowed. Binary redistribution of FPGA RTL allowed with project approval. Source registribution not allowed. Re-licensing not allowed. Design level improvement must be contributed back RTL level improvment must be contributed back High license cost. (Due to the high exposure and possibility for leaks) Hobbyist/Project Developer: FGPA RTL only Binary redistribution allowed. Source registribution not allowed. Re-licensing not allowed. Design level improvement must be contributed back RTL level improvement must be contributed back Cheap license cost (person is effectively a project developer). I haven't really thought through all the differences carefully. Consider this an example of possible license differentiation points. Patrick M _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
