1325323 - Is "Open new private tab" telemetry problematic?

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: General, defect, P1)

Description

[Sebastian Kaspari (:sebastian; :pocmo)]

I just noticed that we have code making sure that the telemetry for "new tab" and "new private tab" is the same and we cannot distinguish between the two:

https://dxr.mozilla.org/mozilla-central/rev/10a5e2fc24e4e24fbe115fb190bcc28df4627ee2/mobile/android/base/java/org/mozilla/gecko/BrowserApp.java#3561-3568

In the Activity Stream code that we are currently developing we send different telemetry based on whether we open a normal or a private tab:

https://dxr.mozilla.org/mozilla-central/rev/10a5e2fc24e4e24fbe115fb190bcc28df4627ee2/mobile/android/base/java/org/mozilla/gecko/home/activitystream/menu/ActivityStreamContextMenu.java#250-260

Is there a reason to not do that? Should we change that part? (Note that we are not shipping Activity Stream yet)

Comment 1

[Sebastian Kaspari (:sebastian; :pocmo)]

'Blame' helped me to track this down to bug 1216265. Unfortunately that bug doesn't describe why we did that.

liuche: Do you know more about this?

Comment 2

[:Grisha Kruglov]

If we decide that this kind of tracking s problematic, we can disable it before it hits the "release", whatever that might mean in A-S's case. The intended purpose of these probes is to see how people are using the new interfaces to inform future iterations, so conceivably we could have obtained enough of that information through a limited-audience, staged release kind of a testing process. How representative that information is of a "real" audience is a good question to ask.

Comment 3

[Chenxia Liu [:liuche]]

Thanks for the heads up, Sebastian! Unfortunately, I don't think we can track that due to consistency in what Mozilla data policies are, and we should remove that probe. I filed bug 1325446 for removing the distinction between PB/normal in that probe.

Here's some background:
When Private Browsing + Tracking Protection came up, there was a big discussion about whether it was okay to track Private Browsing usage. I'm not sure about the outcome of that decision in the context of Tracking Protection (Javaun Moradi would know), but this is a wiki page on PB guidelines [1]. Distinguishing between a normal vs private tab would fall into the Local Privacy + Stealth part of those guidelines.

I'd also point to bug 1204995, where we can send information about whether a *pref* related to PB has been toggled, but that is okay because we are not sending usage data about PB.

There are certainly differing ideas around whether collecting usage data of a feature (that happens to be PB) is okay and good, and I've heard several framings of the question (does an average user feel their privacy is being violated if Firefox sends a ping to Mozilla every time they open a private browsing session? where does respecting users' expectation of privacy measure against collecting the data?), but I guess that's out of the scope of this bug.

[1] https://wiki.mozilla.org/Private_Browsing

Comment 4

[Sebastian Kaspari (:sebastian; :pocmo)]

Thanks!