NSF should recommend open source distribution of software developed through its programs, while also recommending that grantees be aware of different open source license options and requirements at the grantee’s home institution.

Why the shift?  With open source becoming increasingly mainstream, and its collaborative nature more fully understood, it likely comes down to a prefect triad of interests:  Agency good stewardship of public funds; agency frustration with seeing government-funded software code often go unused; and visible public frustration with not fully benefiting from their tax investment in research and development.

While this makes a great deal of sense, a challenge remains with the grant recipient’s institution, where an institutional history of capitalizing on the value of intellectual property may still be the order of the day and open source may be less understood by a university’s legal team.

Here then are a few resources we’ve found for revisiting the research institution’s policy – from rationale to resource, sample policies and practical primers. Have some to share?  Let us know.

• “Why Isn’t There More Open Source in Research?“ Josef Weinbub, Institute for Microelectronics, Technische Universität Wien
• “The Need for Open Source Software in Machine Learning” from the Journal for Machine Learning Research (download journal paper)
• “An Open Source Platform for Personal Robots ” Willow Garage (cogent argument contained for releasing research under an unrestrictive license)
• “Open Source Primer for Software Creators” Stanford University’s guidance
• “A Quick Guide to Open Source Software for the Scientist-Programmer“ Morin A, Urban J, Sliz P (2012) courtesy Public Library of Science
• “Why the Public Domain Isn’t a License” by Lawrence Rosen – Making the case that relying on public domain is not good enough for sharing software