[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Free World Licence possible improved structure.

Ross N. Williams <ross@rocksoft.com> wrote:

> I've had an idea for a new Free World Licence structure.
> First, I backtrack to non-contract structure. So it would have a
> structure similar to the GPL. Second, I include a clause like this:
>    You may not copy the Module onto a non-free platform.
>    However, as an exception, you may copy the Module onto a
>    non-free platform for the purpose of transmitting the
>    Module to a free platform or for browsing the source code,
>    provided that you do not execute the Module on the non-free
>    platform, and provided that you take reasonable steps to
>    ensure that the Module is not executed on the non-free
>    platform.
> ...
> So, for example, a company might sell a CD-ROM containing software
> released under this licence. Anyone buying the CD-ROM would own
> a legally obtained copy of the software - ON THE CD-ROM! But to use
> it, they would have to copy the software into their computer
> and this is the point when the clause kicks in.
> I'm not a lawyer, so this mechanism might not be sound, but if it is
> sound, it seems to me that this structure could work. Its advantages are:
>    * Does not require the invocation of a contract.
>    * No messy click wrap mechanisms.
>    * Allows distribution using free and non-free platforms.
>    * Allows non-free platform users to browse the source code.

A first impression is that this seems to me a much better idea than the
previous version. At second thought, some modification is needed because a
copy of the Module resides on a piece of hardware, not on a platform. I
believe this is also at the heart of Tomasz Wegrzanowski's objections. The
licence being a legal rather than technical mechanism, I think it is
permissible to address the motives for making a copy (in fact the text above
already does do): one could say "You may not copy the Module with the
_purpose_ of making it available for execution on or compilation for execution
on a non-free platform." One could add, as above, that having made the copy
reasonable measures should be taken to prevent execution or compilation on
non-free platforms; this would avoid weaseling defenses such as "Hey, when I
copied the Module to my dual-platform machine it was with the intention to run
it from the free platform only, but later I got to like it so much that I ran
it from the other platform too, but without making new copies."

An advantage of addressing the purpose of making a copy is that it
automatically defeats all kinds of schemes intended to get to execute on a
non-free platform without violating the copyright licence: if you have a copy
that is running on a non-free platform, then this did not happen by accident,
so without knowing the details one can safely conclude that at some point a
copy was made with the purpose of making this execution possible. Another
advantage is that it unambiguously puts distributors in the clear as long as
they refrain from compiling or otherwise configuring the software specifically
for non-free platforms.

Some loopholes may still be found, but I think they are fairly minor. If a way
can be found to directly execute a copy on a platform-independent medium
without the need to compile or make a local copy (say directly from a mounted
CD-ROM; this is more likely for scripts than for C-code), then the software
might be legally used on a non-free platform in this manner. This is probably
sufficiently inconvenient that the author of the software need not worry too
much about it; it does however provide a good reason for forbidding
compilation for (rather than on) non-free platforms.

What still gives me most doubts is the probably quite common scenario of a
public ftp/http server, and somebody downloading to a non-free platform. Since
you want this type of distribution to be available for free platforms, and the
server will not be able to tell the difference, the server operator
(distributor) must remain cleared from any blame. But to blame the person
downloading one must firstly construe this act as making a copy (I have been
arguing previously that it is the server that is making the copy) and secondly
arrange that he has a reasonable chance of knowing at that time that such
copying is illegal. As for the first point I think one may regard this as the
server broadcasting the information (since it will respond to any request),
while the person downloading is creating a physical copy from that broadcast
(much like recording a TV program on video); there must be lots of
jurisprudence whether that constitutes a breach of copyright or not. As for
the second point, just because copying copyrighted material is forbidden by
default does not imply one has to assume downloading is illegal until one has
proof of the contrary; since downloading is practically the only service
provided by the server, one is entitled to think this is normally legal. Just
as an experiment I tried to view a copyright notice before downloading a
non-free package (netscape-base-407 4.07-1) from debian by http. I did get the
opportunity (no obligation) to click on "copyright file"; interestingly that
only showed a page telling "No such copyright file", but OK, I suppose the
intention was there. By ftp I think this would be more difficult still. I
don't like the idea of saying the download is illegal but the user will be
forgiven if he opens up the file, reads the notice, and then destroys the copy
before trying to compile it. But I guess there is no real choice if you want
to allow distribution on the Internet, yet invoke copyright law to prevent
execution on non-free platforms.

Marc van Leeuwen

Reply to: