3 ms·
> Mozilla is up front and transparent about what it's doing A statement in a blog that only a tiny percentage of users will read is not at all transparent. Do
by forapurpose 9y ago
> Mozilla is up front and transparent about what it's doing
A statement in a blog that only a tiny percentage of users will read is not at all transparent.
Do users who receive browsers with Cliqz get a clear, effective opt-out warning? Really, it should be opt-in but that doesn't appear to be the case:
The users will receive a version of Firefox with Cliqz recommendations enabled out of the box.
https://blog.mozilla.org/press-uk/2017/10/06/testing-cliqz-in-firefox/ https://blog.mozilla.org/press-uk/2017/10/06/testing-cliqz-i...
- dao- 9y ago> Do users who receive browsers with Cliqz get a clear, effective opt-out warning? Really, it should be opt-in Either way would render the experiment useless (self-selection bias). Obviously if something like this should make it into Firefox proper, that would make something like an opt-out warning a hard requirement.
- Silhouette 9y agoEither way would render the experiment useless (self-selection bias). But that is Mozilla's problem, not their users'. As probably would be the legal position in many places, if news of this got out more widely and regulators started getting involved. It's hard to see how this wouldn't fall foul of even general EU data protection laws today, never mind anything stronger that is specific to Germany.
- dao- 9y ago> But that is Mozilla's problem, not their users'. Sure, I'm just telling you how we arrived at the current situation and why your proposed alternative is a nonstarter for an experiment that tries to figure out if this feature adds user value in the first place.
- daveFNbuck 9y agoIf the warning is a hard requirement after the experiment, running an experiment without the opt-out warning biases your results. Having the warning in the experiment gives you a better picture of what happens to users when you present them the option.
- dao- 9y agoExcept that the question that the experiment wants to answer isn't what happens when you present users with the option. That would only become relevant later. The question right now is whether Cliqz' address bar autocomplete results are superior to the status quo.
- Sylos 9y agoIt would have to be opt-in or the user would have to be clearly informed, if it actually collected personally identifiable information. That's German law. But it doesn't.