News for hackers. Stuff that matters.
Re:Mozilla’s 18-month effort to market without Facebook by skybrian on Thursday October 17, 2019 @02:38PM (#21285084)
Storing things on a Google server isn't the same as sharing the data with Google for other purposes, for both technical and legal reasons.

This really depends on your trust model. Do you trust Google engineers to implement their own permissions correctly? Are you concerned about mistakes, or that they're entirely lawless?

Physical control means you can independently verify that no sharing is done, but it's not the only trust model.

Re:Mozilla’s 18-month effort to market without Facebook by lucideer on Friday October 18, 2019 @04:30AM (#21289604)
Physical control is the only trust model with any kind of practical verifiability, so anything outside of that trust model is essentially blind trust.

This is obviously not a viable approach pragmatically (Mozilla must e.g. trust the hosting provider for their own data, distribution pipelines for their software, etc.), but these are all cases where reasonable precaution can be taken (signing, etc.). Google Analytics' APIs support no such measures.

> Are you concerned ... that they're entirely lawless?

This is a bit of a strawman / hyperbolic question. A company or its employees needn't be "entirely lawless" (whatever this extreme phrase even means), to be outside of regulatory compliance. And the latter is something Google are found to be on a very regular basis.

Re:Mozilla’s 18-month effort to market without Facebook by skybrian on Friday October 18, 2019 @12:09PM (#21292557)
What's an example of Google sharing customer data when their contract says they're not allowed to?