Follow

We do need to get away from the belief that all data collection is inherently bad because it is not. It is possible to collect metrics for software improvement while still respecting the user’s privacy

To survive, FOSS has to steal my data like the adults
abe-winter.github.io/2018/10/0

@LibertyPaulM As someone who has done user tracking and analytics, you don't have to do this. Sure, it'd be cool and shiny, but the infrastructure required to maintain such a program is expensive and temperamental, and the benefit over, say, just asking your users about it or listening to their feedback is essentially 0 except in the case of apps where they have 0 buy-in (like enterprise software, usually the opposite of the model from FOSS)

@LibertyPaulM This is not to mention that most of the analyses that you can accomplish on most user data are only rarely significant in the statistical sense, and almost never causal when you need them to be. Data programs seem to be extremely helpful, and there are big stories in the industry, like google's button color, but almost all of it is pure marketing and hype. It doesn't drive actual user value, and it detracts from all the things that FOSS is actually good at.

@LibertyPaulM I couldn't disagree more. There's always the possibility of opt-in ("send now" for crash reports for example). Most other things could be improved 90% of the way with qualitative instead of quantitative feedback. Do a classic usability test or panel. Don't do opt-out data collection.

@LibertyPaulM what you, what open source projects or what closed source projects intend to do with the data is also just one side of the medal. Once data is there, it's often used for different purposes later. Think of hackers or law enforcement. Don't collect data that's not directly related Me vital to the service/function in question.

@LibertyPaulM I don't know how the "Me" ended up in there. It should have been an "OR".

@LibertyPaulM We will survive. Thanks for the advice though Pete.

@LibertyPaulM
I disagree. in my opinion data gathering is the prologue to making a decision or answering questions. all that data exist in a scope. when too much data is gathered the scope gets too wide and the target of the whole exercise shifts. so no answer, poor decision, pollution of other scopes, ww3 and

Sign in to participate in the conversation
LinuxRocks.Online

Linux Geeks doing what Linux Geeks do..