Erdos Number
My Erdos number is 3: (me -> U. Vishkin -> N. M. Alon -> Paul Erdos).
Exercise
Occasionally, I run. So far, my max distance has been 13.1 miles:
Hochsteins still in academia
Bugs in scientific software
Errors in bioinformatics code lead to generation of incorrect matrices used for sequence alignment, which go undetected for 15 years. Ironically, fixing the matrices worsens software performance (letter in Nature).
Swapped data columns leads to incorrect protein crystal structure publications (as reported in Science, official retraction, Genome biology commentary).
New code can't reproduce results of old code in astrophysics application (official retraction in Nature).
Incorrect computer simulation "reproduces" wrong experimental discovery of sonoluminescent fusion (Computational Science Demands a New Paradigm by Post and Votta)
- Taleyarkhan et al, Science 295(2002), p. 1868
- Shapira, et al, PRL, 89(2002), p.104302.
- Seife, Science 295(2002), p. 1808
- Purdue University research misconduct investigation report (PDF)
Computational predictions wrongly characterized performance of a proposed thermonuclear reactor (Computational Science Demands a New Paradigm by Post and Votta).
Differences in precision across multiple seismic data processing applications (TSE paper, Computational Sci & Eng paper).
Bug in a Perl script results in a retraction (doi:10.1371/journal.pcbi.0030158) of computational biology paper. The retracted paper (doi:10.1371/journal.pcbi.0030051) challenged the utility of a particular metric for evaluating the reliability of phylogenetic trees.