I'm always shocked to read about how academia -- especially scientific fields -- punishes failure of the 'Experiment did not succeed' type.

My 6th grade science teacher was adamant that the scientific method is about testing, logging, and refining... While I get not working doesn't usually make for an interesting paper, isn't it important for scientists to share what doesn't work as often as what does?

Or have I just gotten the wrong impression from a few articles I've seen pop up on HN?