Over the past few weeks I've been spending more and more time considering the lean startup methodology. The methodology is really a codification of the best practices for rapid testing.
More importantly it makes you consider what happens if a test fails
Maintaining control of failure is an important point when working in the tech industry -- the entire field can change rapidly and failure is part of the norm. At Google's scale they must assume that, for a computing task involving hundreds of machines, one of them will die in the midst of the work.
Failure is built into their methodology
The build-measure-learn feedback loop sits as a core principle of the lean startup methodology. To be honest, it's fairly obvious -- but that's not necessarily a bad thing.
If you replace Product with Experiment then you essentially end up with the scientific method.
Repeat ad nauseum.
Interested in saying hi? ^_^