Subject: Re: For Approval: Open Source Hardware License
From: "Chris DiBona" <>
Date: Thu, 13 Sep 2007 14:23:21 -0700

I've been meaning to reply to this for some time. We done some funding
of "open source " hardware development at Google (as opposed to
software funding, which we do all the time, mostly using the apache
license) and so what we do is apply an apache license to released
specifications, circuit diagrams and firmware. This serves our needs
well even though we're forcing a fit of a software license to a
hardware item.

That said, it tell people what we -mean- by release, which is very
important, and we like the patent and trademark language. I'm not sure
we need a seperate license just for hardware, but it would be
interesting to see what people -mean- by hardware license. Most
hardware nowadays is fundamentally a datafile with a circuit diagram,
verilog (yes, I'm a dinosaur) 'code' or other similar electronic
files, with the physical artifact acting only as an end product or in
initial runs as a debugging mechanism for said data files.


On 9/9/07, Russ Nelson <> wrote:
> Dr. David Alan Gilbert writes:
>  > The difficulties I see are to do with requirements to do with the
>  > ability of the end user to replace copies of the code or regenerate it
>  > and also to do with the use of open source hardware descriptions in
>  > projects with commercial code; Some programmable hardware is
>  > one-time-programmable and it costs more to get reporgrammable stuff -
>  >  would you want to force a vendor to make a more expensive system?
> I don't think anybody is suggesting that anybody should be forced to
> do anything.  This is more a matter of the hardware maker making
> claims about their own hardware.
>  > But there again the same is probably true for software - does a
>  > vendor have to use flash rather than a ROM to hold open source derived
>  > software to make it easily changeable?  Does he have to provide
>  > and document the pinouts on the board to reprogram a device?
> Well, that's a whole 'nother can of worms.  Everybody accepts that
> software drivers can be open source even if the hardware documentation
> isn't available.  But what does an "Open Source Hardware License" say
> about the hardware?  That all the details of the hardware are
> documented even if not everything is used?  What if the hardware has
> features that aren't implemented because there is a bug in the
> hardware?
>  > The possibility of damaging a system from downloading modified
>  > firmware to a programmable device is probably significantly higher
>  > than that for software.
> That's the theory behind the winmodem hardware makers' reluctance to
> describe how their hardware works.  Same thing for wireless (wifi)
> hardware -- different countries impose different rules on the use of
> wireless -- but what if the driver is open source and people can
> change things willy-nilly?
>  > There may also be difficulties in deciding what tools you need to
>  > turn your hardware description into something useable - often it
>  > is still a colleciton of commercial tools held together with
>  > scripts and magic for the larger designs.
> I think that falls into the "fully documented hardware" bin.
> --
> --my blog is at   | People have strong opinions
> Crynwr sells support for free software  | PGPok | about economics even though
> 521 Pleasant Valley Rd. | +1 315-323-1241       | they've never studied it.
> Potsdam, NY 13676-3213  |     Sheepdog          | Curious how that is!

Open Source Programs Manager, Google Inc.
Google's Open Source program can be found at
Personal Weblog: