Hacker News Reader

Comment by kenniskrag | Hacker News Reader

Maybe too many power user disabled the telemetry :)

I have telemetry disabled, but I would be happy to answer surveys for them. I've done so multiple times in the past. Surveys used to be the standard method for gathering user feedback, but they've fallen out of favor. Now everyone wants to do everything using telemetry and other automated data gathering. I wish surveys were utilized more often these days. Talk to your users!

I think surveys have fallen out of fashion because enough people have used telemetry or analytics now to know that what people tell you is often at complete odds to their actual behaviour :)

The difference between users self-reporting how they use software and what they actually do when you observe them using it is often massive, even amongst technical users.

Maybe but there are many other methods to figure out user needs and how to implement those needs in the best way.

Which ones (honest question)? I know gathering data with questionnaires. But asking people is a hard problem. "Do you want a faster browser", "do you want more choice in the UI" will lead to a all yes result. If you ask open questions you need much more time to answer the questionnaire and much more time to interpret the data.

Sit down with a user and watch how they use your product. This is a well known method for seeing how the product is being used, and what causes friction. No interpretation of user input required.

Mozilla does do that. It looks like that too doesn't seem to give them reason to preserve compact mode.

(Though to be fair, UX research like that does indeed highlight what causes friction, not necessarily how many people are strongly attached to features like compact mode.)

Field studies and usability testing are some techniques that can be used when you want to discover novel needs and not just minmax existing implementations. Shit is hard because it takes tons of work and you have to have the organizational maturity to accept difficult changes.

If we do a field test of browser usage, do we expect the result to leave anything standing that allows typical power user workflows such as HN readers to work?

Sure, "just" target power users in your tests and have hypotheses at the right level of abstraction.

Yeah, but how do you assign a weighting here if you don't have the telemetry?

Maybe Firefox usage is really like 25% power users, but most of those disable Telemetry, causing the priorities to be all screwed up (See the ALSA example!).

Questionnaires?

Reminds me of this: https://chuttenblog.wordpress.com/2020/11/05/data-science-is...

tl;dr Telemetry indicates ALSA is no longer used. Firefox removes ALSA. Said Firefox version reaches release. Audio completely stops working for all users of some distros...that were yanking out the Telemetry.

OMG, even if the telemetry were accurate, 2-3.5% of their users is a non-negligible number of people to use rip audio from... that's ridiculous!

2.5%-3% of the Linux users, not of the total ones.

From https://data.firefox.com/dashboard/hardware Linux users are about 3% of the userbase, so that's 0.1% of total users.

Jews make up about 0.1% of the world's population. If I were to operate a store and deliberately exclude Jews from being able to shop there, would you not defend them? They can choose to adhere to the teachings of a different prophet just like you can install a different distro.

This has got to be the single most ridiculous comparison I've ever seen on this website.

I'd love to know why. Mozilla has made a change that has no other effect than excluding 0.1% of their userbase. No end users were benefited by the removal of ALSA support, but "only" 0.1% were excluded. I've put 0.1% into perspective.

>No end users were benefited by the removal of ALSA support

Don't be dishonest.

Maintaining things requires time and effort. Time and effort are limited resources.

* Nobody was maintaining ALSA support

* Nobody from the community stepped up and offered to maintain ALSA support to prevent it from being removed

* The ALSA backend poked a massive hole in their sandbox, which hurts users. Rewriting the ALSA backend to make it compatible with sandboxing would have required a lot of time and effort for little payoff. That hurts users by wasting developer time that could have been spent more productively.

* You cannot compare a Free Software project removing support for something that nobody volunteered to maintain to anti-semitic persecution, don't be ridiculous. If Gentoo (or whoever) wants to keep ALSA support then they can carry the patch to do so.

  > Maintaining things requires time and effort. Time and
  > effort are limited resources.

Airlines' provisions of Kosher meals also require time and effort. The meals must be prepared and packaged and handled in a parallel effort to the regular meals, not to mention that they must be counted and accounted for. Yet, those 0.1% of the flying public which require Kosher meals are deemed worthy enough to continue providing support for them.

As they should, telemetry is just an euphemism for spyware, despite what apologists will claim.

Reply on news.ycombinator.com