Re: BCS OS Licence

[Anonymous User]

		  Hi Andrew,

Thanks for your comments, but you have assumed a limited scope for my research which is not correct. Also I think you missed the essence of my findings.

I extracted the full CVS history for all projects hosted on SourceForge and derived a statistical model for the activity performed on them. I looked at total activity, activity per month, etc. I also stripped out dud projects that never had any activity as well as ones which looked like code drops (those that the activity only lasted for less than a month).

Then I seperated GPL from non GPL projects and remodelled the two groups. There was no significant difference between the statistical model which represented GPL and Non GPL projects. A certain level of activity on non GPL licenced projects is equally as likely as GPL. GPL projects are still the most numerous, but no more or less likely to attract a level of activity than non GPL ones. This is an important point when you are a business looking to invest in an open source project (there are many good reasons a commercial organisation would want to do this, IBM being the most active I would say).

Note that Non GPL projects make up almost half of the active projects on SourceForge now. I disagree with your assumption that people choose other licences at random. The GPL is the most widely promoted licence and the obvious choice unless you make a conscious choice to choose another. I think people need a good reason to choose a non GPL licence. I suspect that this 'good reason' is generally the need for certainty. Commercial organisations are fearful of the vague definition of 'derivative works' within the GPL. Most of the other licences simply provide a clearer definition of this term as so commercial organisations know where they stand when using the products. 

Take the Apache foundation as an example, they have a phenomenal record in sustained, successful Open Source development and enjoy almost universal acceptance from both business and the general public. The reason IBM chooses to work with the Apache foundation when releasing projects like Derby (Cloudscape) is, I suspect, down to this clarity in their licence and commercial friendly ethos. I don't see IBM or any other company ripping off Apache due to holes in their licencing ?

I'm not saying I have all the answers as it is a big topic but what was pretty clear from my research was that the GPL is not 'special', the Open Source development model works just as well with other licences. 

There is the looming problem of interoperabilty between the ever growing number of licences and a common Open Source licence would be a definite help. We need to be asking why other licences are becoming so popular and working towards a more inclusive, common licence. I'm not sure the GPL3 will be it as it is likely to be tightly bound into the FSF ideology (which does not seem to me to be very comfortable with commercial organisations participating). This will only lead to reducing significance for the GPL as the Open Source development model becomes pervasive within software development organisations. 

I can understand your mistrust of commercial software development organisations, some of them work from traditional competition models and are fairly ruthless. Microsoft of course being the obvious example as you state. However, I think licences such as the Apache foundation and Mozilla licence are perfectly good for protecting the collaborative efforts of the community ? Companies which 'steal' an open source project, modify it and incorporate it into their own are then cut off from ongoing community support and enhancements. A short term benefit likely to turn into a liablity in the long run (if a company couldn't write it sooner or later they will find they can't maintain it either...). Better to collaborate with the community and reduce costs in the long term.

Lee