Friday, November 11, 2011

Decision tree for scientific programmers in bioinformatics

This is one of the syndromes we're trying to fight in BioAssist...


  1. One reason for this might be that writing code is easier than reading code. See also Joel Spolsky's blog about starting over from scratch:

  2. Thanks for that link, Kees! Never argue with Joel Spolsky, his arguments are wise, and the "I can rewrite this on a Friday afternoon" fallacy definitely plays a role. Our issue, however, is even more profound: in many cases that I address with the flow diagram actually reading the code would not have been needed, let alone rewriting it from scratch.

    What I hear a lot is: "Oh, yeah, that is a nice library. It does 90% of what I need, but it misses xxxxx. That makes it useless for my purpose." Scientifically, this is a /non sequitur/: the second sentence does *not* logically follow from the first. If the library misses xxxxx, you can get xxxxx from another library, or add xxxxx yourself, or call or mail the authors and discuss your needs. Maybe they tell you about another option they have, built for this purpose, or else they may be willing to collaborate with you on adding xxxxx.