mhoye
Everyone _says_ they want choices, but to a first approximation the number of people who change the defaults in any computer program is so close to zero that if the numbers were the only thing that mattered, user choice - provably, measurably - would not matter. But it does matter.
Furthermore: if people depend on your product - not just use but _depend_ on it - then you need to know how that product operates or fails in real-world use.
mhoye
Way, way back in the day, there were two ideological camps at Moz on the topic of software updating.
One camp believed in the idea of positive obligations of user protection; that developers had an obligation to proactively keep users protected from an increasingly dangerous and growing web.
The other camp saw any less-than-consensual changes to somebody's computing environment as a personal violation, akin to a physical violation.
mhoye
This difference of views, decades ago now, culminated with a screaming fistfight in San Jose parking lot.
The "winning" camp of this altercation stayed at Mozilla, and Firefox couldn't auto-update for another... twelve years? Fourteen?
The "loser" left Mozilla and went to Google to become one of the earliest members of the Chrome team.
You don't know what version of Chrome you're using right now. You don't care. It's just 'Chrome'.
mhoye
And let's be 100% real about this, "Chrome auto-updates itself" is probably the most important single development in software security this century, a tectonic change in not just infosec but in how software everywhere is developed and deployed at all.
And they don't worry about the burning tightrope problem at all, because whatever they say, Google doesn't give a single fuck about your privacy beyond protecting your data as an asset.
mhoye
And the combination of those two things means that if you are a big enough, important enough software project, then you can't rely on bug reports or random ragesite comments to tell you when your product is failing. You have to have telemetry. It can't be opt-in because in any scale that matters _opt-anything doesn't exist_.
And if you also care about protecting privacy, then you start walking a burning tightrope.
That's the job.
mhoye
These same conversations happened when Mozilla introduced in-product telemetry; nobody even remembers it now, but the arguments then were exactly the same.
And then, like now, I think that the naive absolutism of free software culture, the refusal to engage with questions like this, are one of the root-cause reasons that the FSF/OSI crowd have sat out this century's most important software engineering conversations.
mhoye
To give you a sense of the ground that's been lost in that absolutism, I'd like to tell you a story about Early Mozilla.
I'm going to elide some names and embellish it a bit, partly because I've only heard myself second- or third-hand and partly because like telling a good story, but as parables go I think it's informative.
It's about automatic in-product updates, before those existed anywhere.
mhoye
If you have a large enough user base that people depend, for whatever personal, safety or security reasons, on your product, then I believe you have a positive obligation to those people to protect them from risks and failures they might never see or understand.
Not because anyone's dumb or incompetent, but because those threat actors make every effort to be invisible to their victims and impossible to understand to defenders.
Alexandre Oliva
Alexandre Oliva
Alexandre Oliva
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.