On Thu, Dec 20, 2001 at 10:56:54PM +0900, Christian Boos wrote:
> On Thu, Dec 20, 2001 at 10:05:21PM +0900, Christian Boos wrote:
> > On licensing issues...
>   :
> > I'll choose the Ruby license, as this choice won't prevent me to bundle
> this
> > piece of software with the rest of our closed source software (that
> doesn't
> > make sense, given what we are selling, but I want to have the option,
> > just in case).
> 
> As long as you are the only author of a software, you are always able
> to distribute it under any license you want. Even if you distributed
> it under the GPL, you may include it in your own proprietary software
> or sell/give the software to someone else under a license which allows
> them to include the source in software which is not free software.
> 
> --------------------
> 
> That's pretty theoretical. Once you give something to the community as GPL,
> you may as well give up the copyright to the FSF :)
> 1) if the software is interesting and is free software, you won't be
>    for too long the "only author"
> 2) enhancements made by other people have to be GPLed
> 3) you won't be able to integrate those enhancement in your non-GPLed
>    version of the software
> 4) you're stuck ... or you have enough manpower to do a "clean room"
>    re-implementation of those extensions
> 
> Point 4) is not an option for me. That was the main reason why Trolltech
> waited so long to adopt the GPL for Qt: Once they felt they were "strong"
> enough to prevent forks by being able to reimplement whatever would show up
> in GPLed code, they did it.

It is not fully theoretical. Some Companies actually develop free
software while watching closely that the only 'real' changes done to
the code are made inside the company. IIRC Kent Beck and Erich Gamma
didn't accept patches for JUnit in the beginning either: They asked
for new test cases instead. (But I do not remeber how 'free' JUnit was
back then, or even how it is now).

Taking this route does limit the usefulness of free software for the
author. But it gives you more control. I wouldn't want to take this
limitation, but it may be a choice, especially for a company.

BTW: Regarding the 'real' changes: It is legally not quite clear how
     much one has to contribute to a project to become one of it's
     authors. Most people wouldn't find it sufficent to simple point
     out that an 'i>15' should rather be an 'i>16'. I once heard from
     the FSF, that if people submit contributions above 20 lines, they
     ask them to either disclaim rights or assign them to the
     FSF.

-- 
marko schulz