Subject: Re: [ppc-mobo] Re: GPL-like hardware design license?
From: Mark Shewmaker <>
Date: Thu, 7 Oct 1999 03:44:52 -0400

On Wed, Oct 06, 1999 at 10:43:54PM -0600, Richard Stallman wrote:
>     with this project, it looks to me that the only sort of protectable
>     intellectual property left would be the copyrightable parts 
>     of the design (schematics and the layouts), which would be suitable
>     for putting under the GPL or LGPL, (or something close.)
> The drawing of a schematic can be copyrighted, but the actual
> circuit graph cannot be.  So anyone who wanted to make a modified
> version of the circuit would just have to redraw it, to get out
> of whatever license you might use.

If it's as clear-cut as you imply, with the copyrights on schematics
so easily worked around, then that's fine with me--it would imply
that circuit designs are effectively as close to permanently being
in the public domain as to make no difference.

They could always incorporate any parts of our circuit designs they wish,
and we could always incorporate any parts of theirs that we wish, for any
definition of "they" and "we".  And it would be practically impossible to
hide the source code, as it were, at least from a skilled and well-funded
advisary.  (It always turns out to be easier and cheaper to examine
and reverse-engineer tamper-proof chips than the proponents ever predict.)

Of course, patents put a severe dent in the utopian paragraphs above.

(But if software worked like that, with the laws of physics practically
mandating the distribution of source code, and the legalities making
copyright encumbrances ineffective, then there would have been no need
of the GPL in the first place, what with it being impossible to prevent
sharing of ideas and implementations.  There'd also have been no need
for any gtk-vs-qt license flamewars.)

> Didn't my article say this?

Yes.  Let me quote the relevant part of your article 
at <>.

|Circuits cannot be copylefted because they cannot be copyrighted.
|Definitions of circuits written in HDL (hardware definition
|languages) can be copylefted, but the copyleft covers only
|the expression of the definition, not the circuit itself.
|Likewise, a drawing or layout of a circuit can be copylefted,
|but this only covers the drawing or layout, not the circuit
|itself.  What this means is that anyone can legally draw the
|same circuit topology in a different-looking way, or write
|a different HDL definition which produces the same circuit.
|Thus, the strength of copyleft when applied to circuits is
|limited. However, copylefting HDL definitions and printed
|circuit layouts may do some good nonetheless. 

You copyright expressions, not ideas, whether the subject
matter is software, hardware, paintings, or novels.  In other
words--I don't see how copyrighting hardware fundamentally changes
the licensing concerns.

A software example is Microsoft copyrighting their implementation
of their SMB protocol, but not the protocol itself.  Then the SAMBA
team copyrighted and GPL'd their implementation of the same protocol.

That's a software version doing what you suggest is doable with
hardware (and implying greater ease of doing so):  "What this means is
that anyone can legally draw the same circuit topology in a
different-looking way, or write a different HDL definition which
produces the same circuit.  Thus, the strength of copyleft when
applied to circuits is limited."

The software analogy would imply, I think, that GPL is a more useful
tool for software that's harder to reimplement, but that it wouldn't
be needed so much if software were close to trivial to reimplement.

I have to admit ignorance in understanding how strong copyright
protections really are on circuit designs, and how easily those
protections can be bypassed by a redesign.  Not knowing the facts
there, I don't dispute your claim, I just say that it doesn't change
the fact that there's a pleasant conclusion re: copyrights whether
you're right or not.

However effectively strong a copyright protection of the expression
of a circuit design is, it would have to lie in a continuum between
the following:

1.  A given expression of the circuit designs is fundamentally
    uncopyrightable.  (Or at least, the copyrights are so easy
    to bypass that they may as well not exist.)


2.  A given expression of the circuit designs is copyrightable, and
    the copyright protection has a lot of strength to it.

If the copyrights work in a way closer to 1, then we're in a world that
enforces public domain and "source" availability in regards to circuit
designs.  If copyrights work in a way closer to (2), then a GPL-type
license protecting the copyrights should work.

So it's either a non-problem as far as copyrights go, or a
solvable problem--good news either way.

Now that that's out of the way, the real things I was wanting to get
across in my previous articles were:

1.  As far as licensing issues go, Copyrights and Patents are
    the things to be concerned about.
2.  Any problem with copyright licensing should turn out to be
    solvable, if it even exists.
3.  I would suggest thinking of copyrights and patents separately
    when trying to think through licensing issues.  (Even though
    you could later design a single license that somehow ties the
    two together.)
4.  (2) implies that Patents are the things to be concerned about.

And here's my ulterior motive:

5.  I've been trying to think about just the patent end of things,
    and I'm trying to put together a GPL-style patent license.
    (So I directed interested people to my site, with a comment
    that I haven't thought through how to graft in an LGPL-style
    patent license.)

So I think the copyright part of the problem is probably solvable
relatively easily, and the patent part the potential land mine,
hence the redirects to my site.  (Which if you look at the site,
gets redirected to the FSB list for discussion, interestingly enough.)

 -Mark Shewmaker