Fundamental issues with open source software development is an essay in First Monday that lists 5 problems with many open source tools. The essay is by Michelle Levesque at U of T and is based on her experience with adapting an open source package. The problems are:
– Poor user interface design
– Poor documentation
– Feature-centric development
– Programed for the programmers
– Religious blindness
She points out how many of these problems also apply to commercial developments – the question is whether any of these are linked to the nature of open source development. She doesn’t quite complete the job of working from characteristics of OS development to the problems demonstrating the inherent strengths and weaknesses in the approach. That is perhaps for a further study.
My question is how this analysis of the problems would apply to Open Research? Can we imagine analogous limitations to open research? Here goes:
– Poor dissemination design – OR projects won’t bother to carefully design the disseminated results both because the results flow out (rather than get published at junctures) and because there isn’t the incentive to use design to sell the publication.
– Poor contextualizaton – OR projects could overwhelm new readers who haven’t followed the project. OR projects wouldn’t bother to place and market results in the proper context.
– Overwhelming detail – OR projects would overwhelm readers with details of the project and not have the benefit of an editor who would force the researchers to balance the work.
– Researching for the self – OR projects would be by a team for the team and could slip into vanity research – “watch me do this, who cares about the results and their applicability”. OR projects could be more about the team than the problem. The medium of OR would be the message.
– OR could become blind to commercially published or supported research. After all we all need excuses not to read research – there is too much around. OR teams could be tempted to only interact with other OR teams.
Note that this adaptation is based on a guess of what OR would be like – I don’t know if we have any examples of completely open research that self-consciously thinks of itself as open. Lets try it while remembering to beware.
(Thanks to Matt P. for the reference.)