Berlin 6 Open Access Conference: Sergey Fomel


Published on

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Berlin 6 Open Access Conference: Sergey Fomel

  1. 1. Reproducible Research Sergey Fomel The University of Texas at Austin
  2. 2. Outline <ul><li>Personal experience </li></ul><ul><li>Why do reproducible research? </li></ul><ul><li>How to do reproducible research? </li></ul><ul><ul><ul><li>Reproducible computational experiments </li></ul></ul></ul>
  3. 3. Personal Experience <ul><li>Jon Claerbout and RR at Stanford </li></ul><ul><li>Madagascar open-source software </li></ul><ul><li>CiSE special issue (Jan-Feb 2009) </li></ul><ul><ul><li>David Donoho et al </li></ul></ul><ul><ul><li>Randall LeVeque </li></ul></ul><ul><ul><li>Roger Ping and Sandrah Eckel </li></ul></ul><ul><ul><li>Victoria Stodden </li></ul></ul>
  4. 4. Jon Claerbout’s Story <ul><li>1987: Sunview experience </li></ul><ul><ul><li>Interactive programs are slavery </li></ul></ul><ul><li>1992: LaTeX + cake </li></ul><ul><ul><li>Rebuilding books by a single command </li></ul></ul>
  5. 5. Reproducible Research at SEP <ul><li>Stanford Exploration Project </li></ul><ul><ul><li>Founded in 1973 </li></ul></ul><ul><ul><li>2 Ph.D. students per year </li></ul></ul><ul><li>Reproducible research </li></ul><ul><ul><li>From CD-ROMs to WWW </li></ul></ul><ul><ul><li>From cake to GNU make </li></ul></ul><ul><ul><li>2001 CiSE paper </li></ul></ul><ul><li>The principal beneficiary is the author </li></ul>
  6. 6. The Madagascar Project <ul><li>Multidimensional data analysis </li></ul><ul><li>Started in 2006 </li></ul><ul><li>Open community </li></ul><ul><li>Open source (GPL) </li></ul><ul><li>Three levels </li></ul><ul><ul><li>Building blocks in C </li></ul></ul><ul><ul><li>Recipes in Python/SCons </li></ul></ul><ul><ul><li>Papers in LaTeX + SCons </li></ul></ul>http://
  7. 7. Personal Experience <ul><li>Jon Claerbout and RR at Stanford </li></ul><ul><li>Madagascar open-source software </li></ul><ul><li>CiSE special issue (Jan-Feb 2009) </li></ul><ul><ul><li>David Donoho et al </li></ul></ul><ul><ul><li>Randall LeVeque </li></ul></ul><ul><ul><li>Roger Ping and Sandrah Eckel </li></ul></ul><ul><ul><li>Victoria Stodden </li></ul></ul>
  8. 8. CiSE Reproducible Research <ul><li>David Donoho, Arian Maleki, Inam Rahman, Morteza Shahram, Victoria Stodden </li></ul><ul><li>15 years of reproducible research in computational harmonic analysis </li></ul><ul><ul><li>MATLAB </li></ul></ul><ul><ul><li>WaveLab: 690 citations </li></ul></ul><ul><ul><li>“ Striving for reproducibility imposes a discipline that leads to better work. ” </li></ul></ul>
  9. 9. CiSE Reproducible Research <ul><li>Randall J. LeVeque </li></ul><ul><li>Python tools for reproducible research on hyperbolic problems </li></ul><ul><ul><li>Fortran + Python </li></ul></ul><ul><ul><li>Clawpack: 7,000 registered users </li></ul></ul><ul><ul><li>“ Scientific and mathematical journals are filled with pretty pictures of computational experiments that the reader has no hope of repeating .” </li></ul></ul>
  10. 10. CiSE Reproducible Research <ul><li>Roger D. Peng and Sandrah P. Eckel </li></ul><ul><li>Distributed reproducible research using cached computations </li></ul><ul><ul><li>R language </li></ul></ul><ul><ul><li>Cacher package </li></ul></ul><ul><ul><li>“ We propose that a modular research approach lends itself more naturally to reproducible results .” </li></ul></ul>
  11. 11. CiSE Reproducible Research <ul><li>Victoria Stodden </li></ul><ul><li>The legal framework for reproducible research in the sciences </li></ul><ul><ul><li>Licensing and copyright </li></ul></ul><ul><ul><li>ORL (Open Research License) </li></ul></ul><ul><ul><li>“ We need a license designed with the needs of computational researchers in mind .” </li></ul></ul>
  12. 12. Why Reproducible? <ul><li>Science is the systematic enterprise of gathering knowledge about the universe and organizing and condensing that knowledge into testable laws and theories. The success and credibility of science are anchored in the willingness of scientists to independent testing and replication by other scientists. This requires the complete and </li></ul><ul><li>open exchange of data, </li></ul><ul><li>procedures and materials . </li></ul>
  13. 13. Open-Source Software <ul><li>“ Abandoning the habit of secrecy in favor of process transparency and peer review was the crucial step by which alchemy became chemistry. In the same way, it is beginning to appear that open-source development may signal </li></ul><ul><li>the long-awaited maturation of </li></ul><ul><li>software development as a </li></ul><ul><li>discipline.” Eric S. Raymond </li></ul>
  14. 14. How to Do Reproducible Computational Experiments? <ul><li>Code attached to published results </li></ul><ul><li>Continuous maintenance </li></ul><ul><li>Previous results used for testing </li></ul><ul><ul><li>Test-driven development </li></ul></ul><ul><li>Lessons from open-source </li></ul><ul><ul><li>Intellectual property </li></ul></ul><ul><ul><li>Community </li></ul></ul>http://