[cc-licenses] Regarding SA and "strong copyleft" question
hancock at anansispaceworks.com
Thu Mar 1 09:24:29 EST 2007
wolfgang wander wrote:
> To overcome some of the problems you have outlined, especially the ones
> of publishing images of different license origins, would it be possible
> to state in a future Share-Alike-like license that
> Aggregations with works of any license are allowed as long as
> a) the aggregation itself (not all the content) is
> released under a SA-like license.
> b) content the 'aggregator' has created and has a copyright for
> (i.e. text, images, soundtracks) is released under a
> SA-like license as well.
> This should enable you to use fair use images with your text, and generally
> aggregate any kind of combinable images but would also require publishers to
> give back to the Free pool in a SA manner.
In spirit, I think this would be okay -- splitting the "parent" and
"sibling" copyleft cases. But I think it might be virtually impossible
to write a license that actually has this effect.
Problems to solve:
1) Clearly the aggregate can't really be By-SA or even By-SA-"like"
unless the license also changes to not be binding on the atoms of
construction. For example, if I compose a work constructed of a
collection of quotes (maybe some of them are by me), the license of the
collection does not bind the quotes themselves. OTOH, it *may* bind
other collections based on it.[*]
2) Making it significant that the "aggregator" has also contributed
"content" to the aggregation is highly suspect. That creates bias
against particular individuals and ties the two acts together in
possibly unexpected ways (for example, if aggregate a bunch of media
together, I might have to avoid including any of my own, because it's
under an incompatible license. OTOH, I am the copyright owner, so I can
choose that. UNLESS, the work has already been contributed to by others
whom I can't contact -- e.g. if it's GFDL or GPL work).
I don't think either complaint is necessarily a show-stopper, but
they're not going to be a cake walk either.
This actually makes a hell of a lot more sense for an open hardware
license, which I'm trying to come up with at the same time as this, so
let me just go ahead and state that case for comparison:
Hardware is made of many different layers of abstraction: chip-process
transistors (which are just patterns of shaped/doped silicon),
combinations of those devices into "IP Cores", combinations of IP Cores
to form complete chip masks, embedding of chips in leaded packages,
combinations of chips (and discrete parts like resistors and capacitors)
into printed circuit boards (PCB), then assembly of PCBs into a chassis
(like your computer), combinations of whole systems (like when you plug
your monitor into your computer).
Now here's the kicker: suppose we want free, copyleft designs for chips,
PCBs, and complete systems. What does the copyleft bind?
* Can you design a free PCB using proprietary chips?
(e.g. this is what the LART was -- one of the first open
* Can you use free chips in a proprietary PCB?
* Can an OEM put an Open Graphics Card into a production, otherwise
* Likewise, can we create a Free PC, even if we have to use some
proprietary expansion cards (say we haven't got a NIC yet).
You can keep going in either direction. For example, can you use a
proprietary chemical composition to implement a free chip fab? (although
that's really going to be under a patent, come to think of it). And can
you use free computers in a non-free robot, car, or train system (or
vice-versa?). Just think of that -- would installing a free PC as a
controller in a metro train system force the release of the train's
plans under a free license?
You can argue that each design represents a creative act of combining a
set of uncreated "atoms" into a "work". Of course, one layer's "works"
are the "atoms" of the layer above, and vice-versa.
But we don't want the copyleft on "works" to bind "atoms", nor do we
want the copyleft on "atoms" to bind "works". Either would be an
unacceptable bar to:
1) Users' freedoms
2) Continued innovation
3) Acceptance of free works
OTOH, we DO want copyleft to bind evolution in each domain. So, for
example, we want any chip design based on our chip design to be free.
Likewise, any PCB based on our PCB should be free (but, for example,
they might change the PCB to accomodate a proprietary video
encode/decode chip that we don't have in our design precisely because it
As it happens, I have to make a proposal about how to approach this
problem this coming week (we've been avoiding facing this for awhile now
;-) ), so this is really interesting to me right now. I'm beginning to
think it is really the same problem as the photo case.
Terry Hancock (hancock at AnansiSpaceworks.com)
Anansi Spaceworks http://www.AnansiSpaceworks.com
More information about the cc-licenses