4 ms·
> you also need to explicitly opt-in Wrong, as far as I see: Looking in my about:config, I see app.shield.optoutstudies.enabled=true browser.onboardin
by acqq 9y ago
> you also need to explicitly opt-in
Wrong, as far as I see: Looking in my about:config, I see
app.shield.optoutstudies.enabled=true
browser.onboarding.shieldstudy.enabled=true
enabled by default. The settings that I've changed from the default are shown in bold. These aren't bold. Those are the defaults. Everybody can check.
That means that the user must actively take steps to disable them, if he knows that they exist and where he can disable them.
Every time the user creates a new profile, and most probably also when he "refreshes" an old one, he has by default the studies allowed.
It's even worse in other aspects: through the UI the "Allow Firefox to install and run studies" can be unchecked but it doesn't change the value of "experiments.enabled" to false in about:config.
Apparently the "experiments" allow Mozilla to install the "experimental" extensions to any user, without him knowing. And these extensions are invisible in the GUI! Even if the user goes to the about:config and sets extensions.ui.experiment.hidden to false, it will be automatically set to true again.
- Feniks 9y agoI speak from experience that a lot of FF users don't even know what about:config is or how to edit it. It all seems sneaky and deliberately obtuse.
- lozenge 9y agoIt looks like experiments are previews of potential new features, like letting you pop a video out of a page https://testpilot.firefox.com/experiments/min-vid?utm_source=support.mozilla.org&utm_medium=sumo-article&utm_campaign=testpilot-secondwave&utm_content=sumo-post-minvid https://testpilot.firefox.com/experiments/min-vid?utm_source... Whereas studies collect usage data.
- acqq 9y agoWhy should the "previews of potential new features" in the form of the extensions be hidden from the user, and even if the user "unhides" them be automatically hidden again?
- Ajedi32 9y agoAre you sure that's what those config options do? I tried looking them up, but they don't seem to be listed in Mozilla's config documentation: http://kb.mozillazine.org/About:config_entries http://kb.mozillazine.org/About:config_entries According to the Wiki page I linked in my previous comment, global settings shouldn't even matter in this case; since each SHIELD study must be opted into on an individual basis. (Or at least, that's how it's _supposed_ to work.) Edit: Looks like the wiki was updated to state that some studies can be opt-out rather than opt-in. This also seems in-line with the documentation for SHIELD, which has a section on opt-out studies: https://normandy.readthedocs.io/en/latest/user/actions/opt-out-study.html https://normandy.readthedocs.io/en/latest/user/actions/opt-o...
- acqq 9y agoYour link in edit part is the answer to your question before the edit: https://normandy.readthedocs.io/en/latest/user/actions/opt-out-study.html https://normandy.readthedocs.io/en/latest/user/actions/opt-o... "opt-out-study: Install a Study Add-on Without Prompting The opt-out-study action installs an add-on, typically one that implements a feature experiment by changing Firefox and measuring how it affects the user." They are obviously the topic of: app.shield.optoutstudies.enabled=true That I mentioned. I see a lot of commenters trying to excuse them. The problem is, people allowed the "studies" because Mozilla claimed that they are "measuring" whatever "to make Firefox better." They never told anybody that they are selling the "studies" functionality which silently installs ("opt-out" not opt in!) to the advertisers. I don't know how anybody can defend such an approach.