As a research team with a strong development activity of open source (model-driven) tools we have faced again and again the challenge of industrializing research tools.
In OSS4MDE 2014 (First Workshop on Open Source Software for Model Driven Engineering) we discussed our views on the problem, specially focusing on the Eclipse ecosystem, in the paper: On Developing Open Source MDE Tools: our Eclipse Stories and Lessons Learned.
If interested, all details below:
Abstract
Tool development has always been a fundamental activity of Software Engineering. Nowadays, open source is changing the way this is done in many organizations. Traditional ways of doing things are progressively enhanced or even sometimes replaced by new organizational schemes, benefiting as much as possible from the properties of open source
(OS). This is especially true in innovative areas such as Model Driven Engineering (MDE) in which new tools are constantly created, developed and disseminated, many of them coming from research teams.
This poses some hard questions: What is the actual impact of OS in terms of tool development? How to best take advantage of OS communities? And what are the opportunities for research teams in this context? Capitalizing on
experiences in developing MDE OS tools on top of the Eclipse platform and its license model, we try to give some insights on these questions in this paper.
Full paper: Download
Presentation
FNR Pearl Chair. Head of the Software Engineering RDI Unit at LIST. Affiliate Professor at University of Luxembourg. More about me.
Couples of moments are missing:
– how to ensure the sustainability of the project in the long run? Would the research interest last forever fueled by the PhDs and post docs?
– is there a business model? Does the open source kill it?
– how to be visible if you enter a “dynamic ecosystem” such as Eclipse?
– do we need or not external contributors to the project? what is the right balance?
I only have an opinion/experience on the “long run sustainability”. For that we propose what we call the industrialization triangle (probably better explained here: https://modeling-languages.com/industrialization-research-tools-atl-case/ ) where the research team “outsources” most of the support and maintenance task to an SME that sees a business opportunity in selling services/consulting around the software