Late Night Linux's latest podcast episode https://latenightlinux.com/late-night-linux-episode-124/ covered the #Audacity #telemetry issue and something one of the hosts said made me think about how little useful user interaction data telemetry really provides...(thread)
He commented that he sometimes changes the UI and because of the awkward design, can't remember how to set it back. He just restores his config from a backup.
Telemetry will never tell you why a user doesn't use a UI element. It cannot teach you what people do not use because they can't find it, don't notice it or aren't aware of its purpose.
And if a project looks at a button and they think "Hardly anyone uses this element, let's demote it". But people don't use it because it's not clear what it does, perhaps the software doesn't have hover-over tips, doesn't have a tutorial, the icon is too small, not obvious, poorly-positioned, or should have been a menu option. Now the button is demoted and even more difficult for us to find - and we're just restoring from a backup instead, we're just making do without.
Hands up if you'd donate five minutes if your favourite #FLOSS software asked you to fill in a survey about how you use the software. Put up the other hand if this method isn't creepy, doesn't add bloat and technical debt, preserves your privacy and has zero chance to exfiltrate personal/confidential data.
@danny I guess with telemetry, it's opt-in to everything or nothing. With a survey, one can leave a question blank.
@proactiveservices Not necessarily. It depends on the survey mechanism, but sure, that would be a plus.
Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.