Re: Non-free postscript code in EPS image
On Tue, Jul 31, 2012 at 05:01:07PM -0700, Steve Langasek wrote:
> On Wed, Aug 01, 2012 at 12:45:27AM +0200, Bernhard R. Link wrote:
> > > even
> > > when portions are copyright other people/entities.
>
> > If there is a hint to distrust what people claim about their work,
> > I see no way how a judge could believe a "But I was told it is" if one
> > did not at least check what hints one got.
>
> > If someone claims he has a license from Adobe, then well, believe him
> > unless you run into some statement from Adobe that they do not give
> > away any licenses like that. If someone just claims it is under a free
> > license but does not even refer to those parts having a different copyright,
> > then it gets unlikely enough in my eyes that one has to assume the default
> > of the law: no permission at all.
>
> This notion of documenting the copyright of every single line of every
> single file is a new development in Debian - and not a healthy one.
Every copyright notice means that there is at least a part copyrighted by the
mentioned copyright holder. It is important that we make sure that every
copyright notice has a corresponding license before distributing the software
via Debian.
> Not
> documenting the copyright of bits that were incorporated by way of a binary
> generation is *normal*, and unless those bits are under the GPL, it's not
> generally a problem - except for passing Debian NEW, which seems to have
> taken copyright auditing ad absurdum in recent years.
Can you give an example of such "copyright auditing ad absurdum" where people
processing the Debian NEW queue rejected the package where you would have
allowed the package to enter Debian ?
>
> So no, I don't think failing to include these copyright statements in the
> documentation is any sort of evidence of a lack of due diligence on the part
> of the upstream.
The issue is that the copyright notices from Adobe have no corresponding
license from Adobe. It is possible that upstream has overlooked that.
>
> > And regarding
> > >> Since Illustrator is frequently used for producing output files
> > >> that are expected to be distributed, it would be reasonable to assume that
> > >> the output is liberally licensed and that whatever license is listed in
> > >> the package is in fact the correct one, with no other license attaching to
> > >> this output.
>
> > have you ever looked at some EULAs?
> > A quick look at http://www.adobe.com/products/eulas/ makes me think this
> > is quite unlikely. (If I read this correctly, some things you are
> > allowed to embed, but only verbatimly and for specific purposes, but the
> > limits are quite absurd and I'm not sure it even gives permission to
> > distribute for all the stuff you find in some postscript files).
>
> I've read EULAs before, sure. But the only EULA that has any bearing on
> this is the EULA for the version of Illustrator upstream used to generate
> the PDF, and that one I have not read. If this issue is important enough to
> you that you want to track down the EULA and verify that the embedded code
> isn't freely distributable/modifiable, that's your prerogative;
If the EULA would make the software free to redistribute then the EULA must be
included in debian/copyright.
> but I stand
> by my statement that *Debian* should not block on such an investigation.
I stand by my statement that Debian must not redistribute software with
copyright notices without corresponding licenses.
Regards,
Bart Martens
Reply to: