Redefining GPL?

[Ben Tilly]

On 11/29/06, Michael Bernstein <webmaven at cox.net> wrote:
> I am aware of a company that recently publicly released code (under the
> GPL) to several web-applications that they no longer consider to be
> competitive advantages.
>
> The twist is, that they have explicitly defined 'distribution' that
> triggers the republication requirement to include publicly hosting a
> modified version of the application. And insist that they are within
> their rights to do so.

I've never thought about this, but this raises an interesting question.

Suppose I write a JavaScript library.  And I GPL it.  You use that in
an application.  A user comes and browses your site, the JavaScript
file is referenced and sent to that user.

What are your obligations now?  I am not a lawyer, but it looks to me
like my copyrighted content was distributed to that user.  Does that
user now need to have a copyright notice, a copy of the license, etc
all distributed in addition?  Reading the GPL like someone with
Asperger's syndrome would, you need to do that.  In the real world,
I'd consider it crazy to require you to do that.  (For a start,
sending you the license slows the site.  Plus there is no way to send
it without either modifying it in some way - like putting it in an
HTML comment - or else breaking the web page.)  What ARE your
obligations?

Anyways looking at this situation, it looks like, depending on the
application, hosting a modified version of the application *could*
count as distribution.  If not, why not?

> I've tried reasoning with them by posting comments on their blog
> (including pointing them at the relevant sections of the FSF's GPL FAQ),
> but the comments were never published. I do know they have read my
> comments though, as they followed up to clarify the 'confusion of some
> people in the open-source community' (ie. we're just confused, and
> they're right).
>
> On top of this, they mangle the GPL in a couple of other ways:
>
>         - They specify their subversion repository as the required means
>         of republishing modifications
>
>         - The LICENSE file has a pre-preamble added before the GPL to
>         define distribution and republishing as described above (I
>         believe this violates the FSF's copyright on the license)

Those are both clear no-nos.

> And finally, their pre-announcement blog-post is using the OSI logo as a
> sort of endorsement.
>
> I'm not interested in outing these folks publicly just yet, but I am
> also clearly not getting through to them (nor do I have any more time to
> try and get them to listen to me). Therefore, I'd like to hand this off
> to someone more authoritative and better equipped to get their attention
> and get them to do the right thing.
>
> Volunteers?

As noted, the FSF would be the right people to handle this.  For
starters, their copyright on the GPL has been violated, and the
document that was modified is clear that modification is not allowed.
People who ignore people commenting on their website frequently pay
more attention when lawyers start talking...

Cheers,
Ben