Open Science, Open Software, and Reproducible Code
a marriage of FOSS and Science
Software has replaced mathematics as the modern language of Science, claimed Edward Seidel, the former director of the National Science Foundation Office of Cyberinfrastructure. However, unlike mathematical formulas, which can be written and read by anyone with enough knowledge in the field, software can be hidden behind black boxes and proprietary walls. A March 2012 article in Nature found that more than 90% of papers published in science journals describing "landmark" breakthroughs in preclinical cancer research, are not reproducible, and are thus just plain wrong.
In this talk Bill Hoffman CTO and founder of Kitware Inc., will talk about the importance open source software, open access publication, and open data play in the advancement of scientific knowledge. The scientific process is currently hindered by closed source software, closed data, and closed scientific research publications. Much of this research is funded by public dollars and governments are starting to move in the right direction by requiring the practice of open science. Moreover, commercial enterprises are recognizing the many compelling business reasons to support open source including agility, quality, and community-driven innovation.
Kitware is a software company that creates and develops open source scientific computing software. The software is used around the world for a variety of purposes including high-performance computing. Our business model is to engage in collaborative R&D, customize our open source software for particular applications, and to provide support for it including training.
In the FOSS community, Kitware is best known for the CMake build tool. CMake and the family of tools for testing and packing software play a key role in creating reproducible software based scientific research. CMake itself came out of a US National Library of Medicine effort to create an open source toolkit for manipulating medical imagery. This software process has been used to create an open access journal where the data and code are automatically tested and available to readers. Open Source software combined with open publishing, open data, and software process provide the framework for a truly reproducible language of science.
The FOSS community and the scientific research community mutually benefit from each others work, and yet often do not appreciate the connection that exists between them. This talk will trace some of the connections from FOSS to scientific research, and describe a future where the two communities are more closely aligned.