2 ms·
After the VW incident I do wonder how many laboratory style tests are actually useful. I can't help but think 'that's not how a human interacts with it!' when I
by Theodores 11y ago
After the VW incident I do wonder how many laboratory style tests are actually useful. I can't help but think 'that's not how a human interacts with it!' when I see the robot arm doing its monotony thing for 5 million or so keypresses. Add in a reality distortion field and you can see how 'Hockey Puck' mice get to see the light of day.
- HeyLaughingBoy 11y agoMost of them. You can't simply test in an uncontrolled environment because then you introduce variables that distort your test results. Sometimes to the point of rendering them meaningless. So you end up doing both: most of the testing under controlled laboratory conditions, and the rest in the "real world" with "real users."