[vsbuffalo]

I agree with him, and have been complaining about the same shit for ages (I work in bioinformatics too). Sadly, biologists don't care. We're treated as the number crunchers. The real problem isn't that we waste computational resources, it's that many biologists download programs, run their data through it, and if it spits out an answer rather than an error, they trust it. Since that program probably has zero unit test coverage, and the results may be fed into pharmaceutical decisions, disease diagnostics, etc, you're basically fucked if something went wrong. Lots of us have said this[0].

Minor quibble: genome assembly is definitely still an open problem that's computationally difficult. So is robust high dimension inference, but that falls more under statistics.

I've wanted to leave at least a dozen times too, for the better pay, for working with programmers that can teach me something, and to not have my work be interrupted by academic politics. But the people pissed at the status quo are the ones that are smart enough to see it's broken and try to fix it, and if we all leave, science is really fucked.

[0] http://www.johndcook.com/blog/2010/10/19/buggy-simulation-co...