I've just written up some first-draft GPL- and LGPL-derived licenses.  I
won't wade too deeply into the "how reciprocal should your license be?"
argument other than to say that I personally like the free software
philosophy enough that I'd like a good GPL-like license option to exist.

The proposed licenses are at the URLs below, and I'd appreciate any
comments anyone might have.

https://github.com/ewa/free-hdl-license/wiki/GPL-like-license
https://github.com/ewa/free-hdl-license/wiki/GNU-LGPL-like-License

Thanks,
Eric

On 12/07/2011 01:26 PM, Julius Baxter wrote:
> On Wed, Dec 7, 2011 at 5:51 PM, Bruce Perens <[email protected]> wrote:
>> On 12/07/2011 07:33 AM, Julius Baxter wrote:
>>
>> The main reason why I think there's problems with having the license apply
>> (become viral) to anything instantiated within in it is due to the fact that
>> synthesis tools will often instantiate blocks within the design at synthesis
>> and then reproduce source for it, such as a gatelevel netlist. Of course
>> there's no way anyone would want to ask their technology library vendor to
>> release those things under the open source license.
>>
>>
>> You don't have to throw out reciprocal licensing (don't call it "viral",
>> that's pejorative) because of the synthesis issue at all. An example of how
>> you get around it is the compiler exception in the GPL. Taken from the GPL2
>> text:
> 
> Hi Bruce,
> 
> Thanks for the suggestion. I agree, viral is a rather pejorative term
> and I probably shouldn't be using it when discussing licensing issues.
> 
>>
>> as a special exception, the source code distributed need not include
>> anything that is normally distributed (in either source or binary form) with
>> the major components (compiler, kernel, and so on) of the operating system
>> on which the executable runs, unless that component itself accompanies the
>> executable.
> 
> To be a little bit pedantic first, I'm not sure this is analogous to
> the situation you're in when synthesising RTL for a particular
> technology. I would say the closest analogy to FPGA is to compiling
> software to run on the bare metal of a computer - you could argue the
> FPGA configuration binary "runs" on the FPGA in the same way that
> binary software manipulates the control and datapath signals of a CPU.
> For ASIC, however, we're dealing with something entirely different
> when it gets to a distributable form (ie. is manufactured) and I'm not
> sure something like this compiler exception could apply at all.
> 
> I worry that the point of a reciprocal license, bounded by the module,
> could be interpreted as applying to the IP of anything that is
> auto-instantiated by the synthesis tools. I take your point that you
> can explicitly state that such things are excepted, but what happens
> in the case where you're working around a bug in the synthesis tool
> and manually instantiating technology-specific cells inside your
> design? In a lot of cases you must manually instantiate your target
> technology's RAM blocks in your source (or route all of the signals
> out to the module boundary - not a pleasant task.) So I feel strongly
> reciprocal licenses will never be abided by users of open hardware
> source.
> 
>>
>> It would be simple enough to create an exception of this form for compilers
>> that synthesize VLSI elements for the purpose of creating a working device
>> from the source code.
>>
>>     Thanks
>>
>>     Bruce
> 
> Cheers,
> 
> Julius
> _______________________________________________
> Openrisc mailing list
> [email protected]
> http://lists.opencores.org/listinfo/openrisc


-- 
Eric W. Anderson                     Electrical and Computer Engineering
[email protected]                          Carnegie Mellon University
phone: +1-412-268-1908                                  Roberts Hall 244

                          PGP key fingerprint:
           D3C5 D6FF EDED 9F1F C36D  53A3 74B7 53A6 3C74 5F12

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to