2. you can opt-out easily.
I fully support things like this, every company out there is using some kind of feedback mechanism to get insight how are customer using their products.
And assurances that no personal data will be collected are worthless, we need guarantees, often even mandated by company policies, regulations or contractual obligations.
The hoopla with command line arguments should be enough to underline the fundamental problem with that approach.
How so? It’s open source. You can verify it yourself.
Here's a question for you to consider: .NET Core runs very well on Linux. Linux is a first-class platform and issues which only affects Linux is as a general rule considered a deal-breaker when making a release.
Do you think Microsoft would put such a focus on ensuring good Linux-compatibility if their telemetry suggested nobody ran .NET Core on Linux?
I don't know for sure, but I'm inclined to believe that this telemetry has been among the things ensuring Linux today is a first-class supported platform, because Microsoft is seeing that .NET Core is now gaining traction in a space where Microsoft traditionally has has very little.
It's worthless in that assurances don't mean a thing, even if they are given with best intentions and to the best of their knowledge. In this case I could easily verify that their assurance was wrong (they even admitted in the github discussion that transmitting command line arguments violates their own privacy policy), but the deed would have been done.
And for the rest: as I said, I'm all for telemetry. If it's clearly communicated and opt-in. Other tools do it (as mentioned in github, e.g. yeoman) and live well with it.
By "this" do you mean opt-out or telemetry? The issue that many people have is the opt-out portion.