Research Works and Datasets

• findOSSLicense: License recommender for open source software findOSSLicense. A test dataset used to evaluate similarity among SourceForge projects is available here
  


• Information extraction from FOSS license texts (FOSS-LTE): Information available in licenses is useful for better understanding their content. The text can be used for extracting information on license terms using various NLP techniques. A dataset of licenses we are using for this purpose is available here, whereas the results of the application of topic modeling techniques using MALLET can be found in these results. The assignment of terms to topics and respective results are also available here.
  


• Open source licensing and compatibility: The abundance of open source licenses makes the compatibility among them an important research challenge for any organization that employees open source software. For this purpose we are studying a number of open source projects from different repositories (e.g., SourceForge, Maven) and with different characteristics (e.g., size, programming language). We have tested a number of license extraction tools based on the OSS-license-dataset-1.0 dataset (consisting of 100 projects).
  


• SPDXCompatTools: Checks violations on Software Package Data Exchange (SPDX) files based on the license compatibility graph presented in this publication.
  


• Extracting developers' expertise: Developers' social networks, such as GitHub, and Q&A websites, such as Stack Overflow, can be a valuable source of information for understanding what software engineers do and which expertise they have. Datasets of the work on expertiss extraction sre available here. The dataset contains developers’ expertise in different programming languages.
  


• Scrum-Agile global survey: this survey conducted in 2012 (as part of the bachelor thesis of M. Christou) demonstrates where Scrum and Agile are standing in 2012. Motivated by this global spread of adaptive software development and our personal experience in a Scrum industrial environment it was interesting to demonstrate where agile and Scrum adoption lies today globally in terms of quantities, discover the success or failure rate of and agile- and Scrum-driven projects, perform a comparison among the results of using agile or Scrum-based techniques and of following traditional development approaches (i.e., heavyweight) and study development aspects relevant today (e.g., team geographical distribution). The survey results are available here.


• Other datasets: Dataset of GitHub repositories with license information