Dear Steve and Darren,

> Hi Darren - thanks for your thoughts.
>
> > Given that FSL has been released under GPL in the first place, it must
> > be very tricky to avoid now violating the original GPL with this new
> > licence.
>
> Not really - of course, you are right that the old release must stay
> covered by the GPL, but there is no licencing-related reason why one
> should stick with an old licence for a new release, as long as one is the
> owner of the software's IP. When the GPL talks about software "derived"
> that is referring to when third parties take your software and re-release
> it, not if the owner comes up with a new version. The GPL makes a
> difference between holders of the copyright and licencees. We don't
> develop on the basis of licencing the software from ourselves (!) but as
> the authors...and it doesn't seem likely that the University would take
> itself to court over GPL issues, amusing though that would be ;-)

I agree with the interpretation. I would also like to emphasize that Steve
cannot (though I doubt he wants to) revoke the GPL licenses for source code
under version 2. This give you the certainty that if you derive anything
from version 2, your code can remain GPL-ed.

> Actually, a far bigger problem than any of this is that no-one seems to be
> able to agree on how to spell licence...

I think most people will agree that the spelling of the license is a
technicality better reserves for lawyers. The more important thing for me
is that everyone understands the "spirit" of the licensing terms and obey
it.


> Wrt MEDx, no, MEDx is not released under GPL but under it's own licencing
> arrangements - as stated in our old licence, we could release FSL to
> specific organisations under any other conditions other than GPL if that
> was separately organised.

That's what I meant by the "spirit" of the license. The old license
arragement do mention it but is not bullet-proof: How do you GPL+other
license etc but everyone understands that Oxford is free to exploit FSL
commercial potential and that if you make any improvement, etc, you have to
make it available.

> > I suspect that you all prefer to get on with the business of fMRI/MRI
> > analysis and this arose out of hassles from the legals dept. of Oxford.
>
> Indeed - we're not terribly interested in the licencing side of this
> ourselves given that neither version of the licence affects all the
> academic users of FSL - as you say, we are more excited about the science.


Unfortunately, licensing is also an important issue that we cannot avoid.
FMRIB needs protection from modification to the source code and
misrepresentation, users of FSL need protection from the fact that FMRIB
might suddenly charge them an exhorbitant fees (see MS License 6 for a good
example). This situation is unlikely, but give everyone the peace of mind
to sleep well at night ;)

Best regards,
Cinly