[Project-managers] editing external codes

[Ted Ralphs]

You're right, legal issues do serve as an impediment, but it's not 
foundation policy that created the situation and there isn't much we can 
do about it. We could ignore it, since no one is going to sue the 
foundation anyway (we don't have any money). However, this would be 
irresponsible, since the result might be law suits against users, 
especially companies who are supporting our efforts, like IBM. In the 
end, we have to pay attention to this stuff if we want to grow and 
continue to be supported by those with deep pockets.

The current situation is a product of legal doctrine, which requires 
owners of IP to aggressively pursue infringement claims, and the 
pervasive use of the GPL by people who may or may not understand the 
consequences of the use of this license. It is a matter of opinion 
whether the supposed incompatibility of many licenses with the GPL is 
real or whether the terms of the GPL would be held up as legal in court 
(personally, I think not). But there is no denying it creates 
difficulties for mixing and matching codes released under different 
licenses. There is also a general lack of knowledge about all of this 
stuff among people releasing open source software, which compounds the 
problem.

With all that said, I'm not sure all of this "defeats the purpose" of 
COIN-OR. Certainly, the situation is better and there is less 
re-invention of the wheel with COIN-OR than there was before. We are 
slowly, but surely building a library of components that can be used 
together freely. I think it's amazing what can be done quickly now that 
wouldn't have even been possible five years ago. Sophisticated packages 
like Bonmin are being built relatively quickly to the betterment of all 
users.

If you don't want to put up with the hassles and are willing to take the 
risk, you can always distribute things yourself, but it is unlikely many 
people will use it in this case, since there is no assurance that the 
code is being distributed legally. The advantage of a foundation like 
COIN is that it can provide some assurances that the pedigree of the 
codes associated with the project has been checked out and is OK. This 
results in more users with the resources to support further development. 
This is how projects like Eclipse and Apache have gained so much 
momentum and have so many resources. Without paying attention to this 
stuff, we'll never get there.

Cheers,

Ted

Matthew Galati wrote:
> Hmm -- as far as I can tell, all this stuff sort of defeats the purpose 
> of collaborative open source projects like COIN-or. These license issues 
> seem to encourage "re-inventing" the wheel - which happens all the time 
> - and clearly slows down the advancement of research.
> 
> With respect to contributing to COIN -- let me play devil's advocate for 
> a moment - here's what I suspect goes through a lot of developer's minds:
> 
> So -- as far as I can tell, either --
>  (1) I write it myself - a terrible waste of time (in some cases),
>  (2) I jump through a bunch of legal hoops to get the combined stuff 
> online,
>  (3) I keep my code to my self (which has bits of GPL and CPL and XPL 
> stuff) and distribute nothing.
> 
> (3) sure sounds attractive
> 
> Matt
> 
> PS: Oh yeah, I better say this too:
>   '' The opinions above are my own, not necessarily my employer's nor an 
> official COIN-OR statement. '' :)

-- 
Dr. Ted Ralphs
Associate Professor
Industrial and Systems Engineering
Lehigh University
(610)758-4784
tkralphs at lehigh.edu
www.lehigh.edu/~tkr2