For Approval: GPLv3

Chris Travers chris at
Thu Aug 30 16:57:42 UTC 2007

Actually, the GPLv3 may be better in terms of preserving additional 
permissions than I had thought.  This means that my OSD #9 concern 
doesn't apply, and that the GPL3 may be more reasonable than a first (or 
even third) reading may suggest.

So far many here seems to think that section 7 additional permission 
removal requires sublicensing.   The argument against an encumbrance 
theory would seem to be that the permissions can be removed without any 
changes being added.  This poses one serious issue in reading the GPL3 
which I want to bring up.

Section 2 of the GPL3 prevents sublicensing,  and section 7 specifies 
that additional permissions which affect the "work as a whole" are read 
as included in the license.  It also states that additional permissions 
may be stated in the license itself.  Thus it seems to me that GPL3 + 
additional permissions cannot be reduced to GPL3 because of the 
prohibition on sublicensing.  It therefore seems possible that 
additional permissions could be very difficult to remove in the absence 
of modification to code.

Draft 1 allowed permissions to be removed when code was modified.  This 
would have had a very different effect because one could have simply 
argued encumbrance rather than sublicensing.

If I add to my LICENSE.TXT a statement that says that a linking 
exception exists for such an interface or file, it seems that nobody 
could remove those except by permission of a copyright owner whose code 
was in that interface (someone who had modified the code).  In this 
case, it would be sublicensing under the permission of a copyright 
holder (not otherwise granted in the license).  However, if the code was 
publicly available, for example, in a public CVS or SVN repository prior 
to the license being changed, people could fork at that point and 
preserve the additional permissions.

This interpretation seems to be inline with the first draft, the 
rationale document for the second draft, and other portions of the GPL3 
as released.

None of this is a bad thing.  The only problem is that the possibility 
of multiple interpretations makes me concerned that it would result in 
legal action.  I have no problem with the approval of this license.  I 
do, however feel that we should *not* list the GPL2 as deprecated, 
superceded, or otherwise out of date.

Best Wishes,
Chris Travers
-------------- next part --------------
A non-text attachment was scrubbed...
Name: chris.vcf
Type: text/x-vcard
Size: 171 bytes
Desc: not available
URL: <>

More information about the License-discuss mailing list