3 ms·
When I write a test, I intentionally make the condition fail to validate that the test works like I think it does. The experiment described here uses new equip
by jniedrauer 8y ago
When I write a test, I intentionally make the condition fail to validate that the test works like I think it does.
The experiment described here uses new equipment that presumably has never been used for this purpose before. But I don't see any mention of efforts to verify that it can actually be used to detect black holes. Hopefully the remaining 10 days of their experiment can confirm that it works...
- maxander 8y agoThe “Hyper Suprime-cam” instrument came online in 2012, so it’s not new, and was presumably calibrated against well-understood targets; so its light-detection capabilities have been well tested. The physics of gravitational lensing is also well understood and verified, so they know what light they’re looking for. If you can test each part of a system independently, and the parts connect in a reasonably trivial fashion, it’s about as good as testing the whole.
- Chilinot 8y ago> If you can test each part of a system independently, and the parts connect in a reasonably trivial fashion, it’s about as good as testing the whole. Not really, which is why integration tests are so important.
- scriptkiddy 8y agoDepends. In functional programming, if all components of a pipeline are tested independently, you can all but guarantee that the pipeline as a whole will work as intended. That said, software testing isn't really an apt analogy now that I give it some thought. Testing real world equipment has to contend with a lot more edge cases and environmental factors.