Christian,
It seems that people are well enough informed by the field studies that our team runs to want us to continue to run them. In fact, demand has sky-rocketed both within and outside of the Wikimedia Foundation. You hold a minority opinion that testing software in the field is unnecessary. Yet, field tests are considered a best-practice and have become a critical part of our strategy for minimizing the disruption (and maximizing the benefits) of software changes. I don't think that many people would appreciate your proposed strategy of releasing the software and waiting for people to complain. Given how difficult it is to develop good user-facing software, it's likely that every major deployment would be disruptive if we adopted that strategy. I can speak for a few disruptions that my research helped prevent and some opportunities that it helped us explore.
Allow me to share a specific example. In this study[1], we found that telling anonymous editors to register dropped their productivity by 25%. Yet we didn't identify substantial issues in user testing. If we had not run this field experiment, we might have deployed the change thinking that we were improving Wikipedia when we were really driving good editors away. During the experiment, we received no substantial negative feedback
Really, what I want to say is this: If you want to improve privacy protections, I am your ally. We're merely disagreeing about whether it is good to assume that DNT means something it wasn't intended to mean or not. However, when you say that my work has no value, it's hard to talk to you productively because, honestly, I don't think your opinion is well-informed.
-Aaron