[gdal-dev] change in ECW license agreement

[Tim Bowden]

On Thu, 2008-01-03 at 15:08 -0700, Brent Fraser wrote:
> Interesting problem.
> 
> The "old" GNU license says "This Agreement may only be
> modified in writing signed by authorized representatives of
> ERM.", but are we bound by those new license changes if we
> use only the "old" source?

I'm not trained in the art of law, but my limited understanding is
unless the old license agreement makes provision for being superseded by
a more recently published one (which from memory it doesn't) then so
long as you are using the source obtained under that original license it
still holds.

> 
> I say No, but I'm not a lawyer.  If the answer is Yes, then
> I think the open source community needs to protect itself by
> ensuring licenses for donated source code are granted in
> perpetuity.
> 
> And both the old and new licenses are vague about the rights
> of end-users who are using open-source apps to create ecw
> files for sale.  Because I profit from the sale of data in
> ecw format (but not from the sale of programs), do I need a
> commercial ecw library license?  Maybe.

I tried parsing the old ecw license some time ago for a project I was on
(commercial implementation of open source) and the response of ERM was
that they were aware of the difficulties in parsing the license (as far
as I was concerned it was vague & internally contradictory).  The
problem they had was that there were some third party patents they were
relying on, so could not use an OSI license (good to see the software
patent system is helping innovation again!), and that if the ecw lib was
being used to make money, they wanted their cut.  As far as I'm
concerned, it's not even close to being free software, even if it does
have source available for review.

> 
> Perhaps we need a Google summer(s?)-of-code project to
> produce a high performance FOSS JPEG 2000 library...
> 

libjasper for jpeg2000? What about ecw?  There's lots of data in that
format already existing that needs to be got at. I'm no expert on these
things, but wouldn't development of a free lib take a "clean room"
implementation to avoid using the "methods and concepts" (gee, thanks
SCO!) from existing patents used in the current non-free lib?  Has the
ecw format been published (apart from reading the existing source code),
or would that need to be "reverse engineered" as well?

> Brent Fraser
> GeoAnalytic Inc.
> Calgary, Alberta

Tim Bowden