FDL again, was: My concerns about GPLv3 process

Frank Heckenbach frank at g-n-u.de
Sun Feb 12 13:26:51 UTC 2006

Alfred M. Szmidt wrote:

>    Now you (Alessandro) described how even FDL without invariant
>    sections and cover texts can be harmful (because you can't prevent
>    others from adding some).
> None of what Alessandro described is harmful, it is explcitly allowed
> by the license.

That's why it's harmful. If it wasn't allowed by the license, it
could simply be prohibited by enforcing the license. If you don't
think the outcome (as I described in an extreme form) is harmful,
then we just have to disagree. I think it's harmful, so I don't like
the FDL.

>    Now the FSF can, of course, by the FDL, include the new content
>    into their distribution, but not without also adding the new
>    invariant sections. What would they do?
> Same thing they would do if they wanted to include GPLed code into
> Emacs: require a copyright transfer.

They cannot require it. They can only ask for it, and the "evil"
author can just decline the request.

> This is required for GNU
> projects, and should be required for _any_ and _all_ projects.

This is a requirement the FSF puts on itself before accepting
contributions. It's not a requirement they can put on the
contributors, because the FSF does not have special powers by the
license (unlike NPL and such), which is a good thing. (I know you
didn't say otherwise, but if you think so, this is probably a
misconception on your part.)

So in my scenario, as they couldn't fulfill their requirement,
they'd have to reject the contributions in their distribution, but
they couldn't stop someone else from distributing them (according to
the freedoms given by the licsense). That's what I called the "most
likely outcome" and described with its possible worst consequences.
Again, if you don't see a problem with those consequences, then the
FDL might be ok for you.

> If you wish to have a way to enforce the copyright, then you will need
> to have copyright papers.  Otherwise, you cannot enforce it at all, it
> really doesn't matter what license you are using.  A judge would love
> to hear "Sorry your Honour, but I couldn't gather the copyright
> holders to actually sue the accused'.
> This is a misconception you, and others have shown repatedly when it
> comes to copyleft, that unless you have a single copyright holder (or
> a very small number), it is impractible to enforce the license.  And
> if you cannot enforce the license, the license looses its charm.
> In theory, anyone can go and make Linux a non-free program, since it
> is simply impossible to enforce the license there.

Back up this claim.

I cannot see why a copyright holder of only part of the code cannot
enforce this copyright on his part. E.g., Linus Torvalds alone could
certainly sue someone would did what you said and terminate that
person's GPL rights (according to paragraph 4). Then that person
would have no rights anymore to use a Linux version that contains
any code from Linus, and might also be liable for copyright
infringement or damages, even if only partly for Linus' part of the
code. Unless they'd replace all of Linus' code, they couldn't use
Linux at all then. (And if they do, the next major contributor might
jump in and sue for his part of the code, etc.) If only a small
contributor is ready to sue, someone might be able to step around it
(practically, not legally), but as soon as a major contributor is
ready to sue, I don't see how they could. So FSF's collecting of
copyright assignments is not a necessity for enforcing the license,
but an additional safety to them (for the very unlikely case that
all the major contributors disappear or, perhaps more likely, are
unwilling to or do not have the resources to take action
themselves). IANAL.


Frank Heckenbach, frank at g-n-u.de
GnuPG and PGP keys: http://fjf.gnu.de/plan (7977168E)

More information about the Discussion mailing list