On the Variability of the BSD and MIT Licenses

From AcaWiki
Jump to: navigation, search

Citation: Trevor Maryka, Daniel M. German, Germán Poo-Caamaño (2015) On the Variability of the BSD and MIT Licenses. Open Source Systems: Adoption and Impact: 11th IFIP WG 2.13 International Conference, OSS 2015, Florence, Italy, May 16-17, 2015, Proceedings (RSS)
DOI (original publisher): 10.1007/978-3-319-17837-0_14
Semantic Scholar (metadata): 10.1007/978-3-319-17837-0_14
Sci-Hub (fulltext): 10.1007/978-3-319-17837-0_14
Internet Archive Scholar (search for fulltext): On the Variability of the BSD and MIT Licenses
Tagged:

Summary

BSD and MIT are popular template licenses: they have to be modified to be properly used.

Contributions (quote):

a) An empirical study documenting how the MIT and BSD are modified in practice in source code of Debian software packages;
b) Analysis of this variability., and
c) Recommendations for the SPDX Group on how to address the variability of these licenses in their future templates.

"Ninka was run on all the files of [Debian Linux Distribution, version 6], encompassing more than 1.3 million source files contained in 10,014 projects. Ninka identified 42,653 licenses in the BSD family, and 28,205 licenses in the MIT/X11 family. The variability analysis of these licenses worked in a top-down manner, by first identifying the largest, most commonly used variations, working down to smaller, less commonly used variations to the point where further analysis of variability would be negligible."

Authors provide summary of variability of MIT and BSD licenses, including tree visualization of specific phrases and words changed, added, or removed.

85% of BSD licenses found exact matches for SPDX cataloged versions. Variation most often found in non-endorsement sentence which contains a template field, perhaps encouraging other changes. MIT exhibited more variation, with majority not an exact match, most commonly treating parts of the license as template fields.

Authors recommend for SPDX:

- treating certain words in MIT de facto used as template fields to be treated as such - treating certain words as equivalent (eg contributors, co-contributors, co contributors) - some significant MIT variations should be added to the SPDX license list

Data for this paper is at http://turingmachine.org/2015/mit-bsd/