Hi Bob and all, On Wed, 2007-11-07 at 18:49 +0000, Bob Tarling wrote:
Michiel said: > Was there any negative influence from the forking of Poseidon on ArgoUML? > There was (AFAIK) never any code flowing back to ArgoUML, exactly like if we > would have prevented the fork. The negative influence on ArgoUML was the loss of developers to ArgoUML and the lack of fixes and enhancements back to ArgoUML from the organisation. Imagine if ArgoUML had had a stronger licence at this time of the split how much different our progress would have been. I completely agree with Bob. The loss of key developers and their knowledge and expertise had a very negative impact on ArgoUML Note that I never had a problem with those guys wanting to make a business based on ArgoUML. However, the result of their fork was that ArgoUML progress was severely hampered, IMHO. Soon after they left, they made significant infrastructure changes in their commercial fork they chose not share with this community. In particular, they quickly migrated from the old Novosoft UML library based on UML 1.3 to NetBeans MDR and UML 1.4. If I am not mistaken, it took ArgoUML at least 2 years to catch up with the Poseidon fork in this respect. Imagine where this project would be if all the work invested in migrating ArgoUML to MDR was applied to other important development tasks! It was a huge duplication of effort that could have been avoided with a different license. During that time, I knew it was highly unlikely that they would share their MDR work with us. However, I did make a request of them to just share their UML 1.3 to UML 1.4 model conversion tool. A very isolated piece of functionality that they could have shared with us with no negative impact on the competitiveness of their product. Needless to say, I got no response from Gentleware. The same problem is repeatiing itself with UML2 and EMF. I think Gentleware has been supporting UML2 with EMF for over a year. ArgoUML developers will have to duplicate the same work, even the tools for migrating models from UML 1.4 to UML2. If anybody has an idea of how many person months have been spent by ArgoUML developers working on MDR and EMF related tasks, I would love to hear it. I have no idea what the people at Gentleware think about this issue. What we do know is that Gentleware has no financial or legal incentive to share their improvements to their ArgoUML fork with the ArgoUML community. We can't fix that in hindsight but we can prevent it from happening again. But it seems sticking to BSD will stop at least one commercial developer from commiting code back to us so I think we do need to tackle that by a licence change. I have already stated my license preference. However, I want everyone to know that I still want to work with ArgoUML, regardless of what the community decides to do with the license. Also, my company is small and currently unprofitable. I am optimistic about generating good revenue next year. If that happens, then I will seriously consider paying to develop some features for ArgoUML. However, please don't base your license decision on my speculations on what my company may do in the future.. cheers, roy �
