i really like this idea but feel conflicted. the privacy conscious side of me loves this… but having worked for early dev-tools startups, without telemetry we’d have had no idea how folks were using our tools or how some experiments played out.
without question everything should provide opt-out (or even better, opt-in) for telemetry.
is there middle ground for standardized opt-out vars without it being all or nothing?
i took a look at the prs and issues they linked. i didn’t realize most of this was 6 years old. additionally, the folks submitting the prs seemed ill equipped to make a convincing case for why anyone should be among the first to adopt a nob-standard standard.
even then, one maintainer makes an excellent suggestion: give folks a list of the specific vars to set for various tools, and the pr author shoots it down as too much work.
the approach to this reasonably good idea feels flawed and effectively dead.
i really like this idea but feel conflicted. the privacy conscious side of me loves this… but having worked for early dev-tools startups, without telemetry we’d have had no idea how folks were using our tools or how some experiments played out.
without question everything should provide opt-out (or even better, opt-in) for telemetry.
is there middle ground for standardized opt-out vars without it being all or nothing?
maybe DO_NOT_TRACK_<identifier>?
i took a look at the prs and issues they linked. i didn’t realize most of this was 6 years old. additionally, the folks submitting the prs seemed ill equipped to make a convincing case for why anyone should be among the first to adopt a nob-standard standard.
even then, one maintainer makes an excellent suggestion: give folks a list of the specific vars to set for various tools, and the pr author shoots it down as too much work.
the approach to this reasonably good idea feels flawed and effectively dead.
6 years and about 6 prs and most went nowhere.