It does sort of hinge on the highly suspect assumption that usefulness is correlated with use. An obvious counter-example to this is something like a fire-extinguisher, which will in the ideal case just sit on a wall until it's use-by date passes and then it's discarded having never been used; or on the flip side, an incredibly byzantine workflow that could be reduced to something much simpler will appear important and useful.
Even without these edge cases, interpreting statistics is really hard. Like people with PhDs who have studied these things for years still get them wrong all the time.
What ends up happening more often than not is it's used as a tool to quiet the critics when pushing through unpopular changes.
There's a lot of very simple questions you can answer very reliably, too, like "what proportion of the users are still using a certain compatibility flag".
Interpreting data is very hard.
Funny, I’d assume the exact opposite. After all much of the understanding of privacy and statistics at scale was developed after 1980.
More than that, the interesting stats may be not even around user-visible features, but around internal mechanisms, like some cache hit rate, or how often is some branch in the compiler invoked.
As long as stats are clearly inspectable, reasonably anonymized, and are opt-out, I'd be fine with sending them.
See for example Mozilla's bizarre decision to remove the ability to change the override the character encoding of a webpage with some half-baked detector.
Meanwhile, I'll decide what I must or must not adapt to, thank you.
But, that costs money: folks don't generally like to do surveys for free, don't like to come to your site (or have you come to theirs) for free, don't like to participate in tests for free... and the caliber of person who can design and orchestrate all that is _notably_ expensive.
So, companies have discovered that it's far, far cheaper to shove remote eyes into their products. No need to schedule user observation sessions, pay for transportation and participation, have to bother with hiring people who are capable of coming up with a good set of questions and scenarios to walk through.
This actually happens sometimes: when developing the demo ".kkrieger", a first-person 3D shooter in 96 KiB, demogroup theprodukkt tried to shrink it down to get it under the 96 KiB wire. One of the tricks they used was using a profiler to identify code sections that were never reached and could be removed. One of the sections they removed was the handler for the up arrow key in the main menu, simply because the test player never pressed up in the menu.
If you think that Google or another large software organization won't misuse telemetry by cutting or neglecting important but infrequently used functionality to hit some KPI... have you ever worked in a large software organization?
All stats can be deanonymized. The more data you make available, the more you identify yourself. I do not need software I use stealthily tying up bandwidth by "phoning home" with data about me. It is simultaneously betrayal and resource theft. If I wanted to contribute to the improvement of the software, I'd file a bug report.
But since you clearly don't like telemetry, you should have a way to reliably switch it off. Here we are on the same page: there must be a well-documented and easy way to switch any telemetry off.
Telemetry should be opt-in, if it's provided at all.
Amen
Sometimes some pretty niche feature, that is used by one individual that is loud, gets more attention than a feature, used by thousands or millions that are just silent users.
I disagree, I think it's a very contested topic, with lots of discussion whenever it's raised here, with either side possibly being a vocal minority.
I respect that. I don't agree with the decision, but it was made with understanding and thought.
In ethics, there is no right or wrong answers (mostly), just right and wrong methodologies. If you go the pragmatic way, you'd argue that the benefits of telemetry are greater than the downsides and implement it. If you go Kant's way, you would already have a maxim (either "never invade privacy" or "prioritize technical benefits regardless of the users" in this case) and act according to that maxim regardless of the situation. If you go the intent way, all that matters is whether your intent for the action is good or bad, in contrast if you go the outcome way, all that matters is the outcome regardless of the intent or the methods.
These are all "valid" ways to discuss an ethical dilemma. However, one must always acknowledge the dilemma. This industry, especially big tech, seems to ignore this quite often, mostly because it's very easy to see people as "just numbers" when you don't see them directly. Don't even get me started on lawmakers who are also ignoring this whole issue. Many standard practices in this industry would be straight up illegal in lots of other areas, especially where there is face-to-face contact.
Turns out, my straw man can't read.
Because that's how OP framed it. There is an ethical way to track how your app is used, with user consent.
Furthermore, looking at the asciinema from OctoSQL, users apparently have to edit their profile files on every single machine they intend to use Go on, then remember to verify that the profile applies and has no typos, make sure to never ever accidentally run the program in a context where environment variables aren't respected (would you remember to always use `sudo -E` if you need to use the tool as another user for example?). The danger seems extremely high that the user would, at some point, accidentally run the command without the env var set, even if they were technically proficient and did their best to opt out.
This is not how consent works.
First, making network requests when downloading packages is necessary for the tool to function and unavoidable. People who care about this will be using a VPN of some kind. It's just how the Internet works. But telemetry is something the tool author is choosing to add, not something that's necessary due to the architecture of our computing infrastructure.
Second, the Go telemetry would apparently create a unique, persistent user ID. Normal Internet use doesn't, there's just the IP address which is different from location to location, shared by a bunch of people behind the NAT, and can be masked using common tools.
And yeah, I know this is "anonymised"... but if you have one user ID which uses Go sometimes with an IP address from a particular apartment complex land sometimes from a particular office space, finding out which individual that user ID belongs to is trivial.
It's technically not unavoidable. The Go authors could have made use of the proxy opt-in rather than opt-out, making the tool less usable as a result. A similar argument applies here, I think.
> Second, the Go telemetry would apparently create a unique, persistent user ID
Where did you see this? I scanned through the "Telemetry Design" article reasonably carefully and couldn't find any mention of this concept, and the type definition for the posted JSON (the `Report` type) doesn't seem to include any such user ID.
In the end, ISTM that you're not complaining about something that actually affects your privacy in any way, but just the _idea_ of telemetry. Is that really something worth taking such a hardline stance on?
Frankly, that crap should be expunged from the Go toolchain as well. :(
The distinction is between "What I do with my computer is none of your business unless I choose to make it your business" versus "What I do with my computer is your business unless I choose to not make it your business".
It's insane that we are still having to justify privacy as a default, or that people continue to rationalize away the concerns.
Yeah, maybe if it's opt-in they won't have much telemetry data. Perhaps, in fact, it would not be much better than having no data at all. That will make some things harder. Major bummer. If it was easy to do the right thing, more companies would do it.
Half of the HN is in love with Chrome and nearly all are on Gmail.
The relentless drive towards the erosion of privacy powered by free carrots worked.
There is now a whole generation of people that sees all this as a norm.
The "there shouldn't be any" argument just seems so entitled with there being such demand / reasons to do so. I applaud everyone trying to find ways to satisfy both sides, as done by the original article.
"What I do on my machine is none of your business, unless I decide it's your business" is a strong argument, presented multiple times in multiple places, and I haven't seen any rebuttal by any Go representative.
You can ignore the community and do whatever is good for business, but it's hypocritical to pretend you're a "community-driven" project if you do that.
Opt-out is generally rejected by European privacy laws.
...where personal data is involved.
It strikes me that this proposal goes to considerable lengths to avoid collecting anything that could be considered personal data.
https://gdpr.eu/eu-gdpr-personal-data/
You have to be very careful to do it properly.
Storing IP addresses in logs means that you are now responsible for them, yes. Drop them out of your logs, and you're perfectly fine.
So if you want that information, find a way to ask the user first. If you can give a good and understandable explanation on how the information is useful, the users might give their consent happily.
If you're building a website, I think it's fairly reasonable for you to store my IP. That's inside my expected privacy loss when dealing with a remote party. I have to connect to your computer, much like i have to physically walk into a store. I don't mind you remembering that I was there. Running a compiler on the other hand feels more "private" to me somehow. My expectation when using a compiler is that it won't send anything to anyone, because why would it?
In general I think our industry is starved for relevant and foundational ethics research, outside of the FSF at least.
You make a good point.
As a for instance from a popular Mac-based package manager that (unexpectedly for many) defaults to telemetry from your CLI:
`brew analytics off` is not hard to type after installing homebrew, but the installation text doesn't mention that; instead it points to a web page you have to read about how wonderful the analytics are before eventually finding the incantation:
https://docs.brew.sh/Analytics
I wonder how many people care enough to click that link, read all the "analytics are actually good for you" copy, and then change their mind to leave it on. I'm guessing almost zero?
But perhaps most users won't cut and paste the link, where if it just suggested `brew analytics off` many users would type it.
Then don't do telemetry at all.
Google doesn't have a right to data.
That's too bad. Guess you don't get any telemetry data if you want to develop ethical open software.
The answer isn't to bend your ethics.
If a take a dollar from everyone but it's opt out, that's just theft with extra steps.
If I make it opt in, nobody is going to give me the dollar, but that doesn't make opt out morally justifiable.
Developers can see that a specific crash is being hit by 1% of their userbase and then check the logs to see what went wrong and where the cash happened. The fix is made the program is indeed better.
Crashes, common failures, UI/UX friction points, avarage usage patterns - all can be used to prioritize work to take care of things that have the biggest impact.
Is there an example of a program that was crap, implemented telemetry and then got better afterwards? (and of course controlled for factors that might have improved the program anyway)
I mean since telemetry advocates are so into how useful data is, surely they must have data on whether telemetry itself works?
That option is a joke. The real alternative is rust - or any non-corporate platform that isn't gonna pull these kind of stunts.
Opt-out telemetry won't work when people send false data to the servers.
This is a meaningless point.
"If we ask everyone is going to say no, so we will steal it unless someone tells us not to"
Is opt-out telemetry unethical ... depends. If you use it in a privacy preserving way no, if you spy on your Users, sell the data for money or advertising obviously it is unethical.
The hard truth is, nobody reads the manual. Opt in telemetry is often a minority, and you then work with niche data for a minority that influences your development in certain ways.
> if you spy on your Users
In my opinion, any data collection about me or my machines that occurs without my active informed consent is "spying". This is my fundamental problem with opt-out mechanisms. They do not indicate or imply that active consent was obtained.
Would that be enough for you?
In short, unless you're a Windows user there are so __very__ many ways to install software that aren't "Go to the project home page, download a generic install binary, run that binary with world-write permissions.". Aside from very small-time projects, I can't think of the last time I used an officially-maintained install script that I got from the project's servers to install something.
Why, in OSS, do you care about the users that do not care? If the users truly care they should buy support, or at least enable telemetry.
If someone complains about feature getting removed - tell them to enable telemetry next time. Maybe people who want so much privacy shouldn't be expecting free support for features they want.
Maybe you're worried about people like your grandma installing Linux. Then this should apply on the distribution level - there should be an opt-in setting for telemetry, that enables it in various programs being distributed (at the discretion of the package maintainers), so that user doesn't have to opt-in individually.
Being opt-in will also make it compliant with EU data privacy laws.
This is not true. You know it and you are being coy about it.
If your "easy way to disable it" was a simple question next to that unexpected notification displayed once in response to a completely unrelated action - "Would you like to keep telemetry on?" - you bet you'd have massive opt out rates. Nobody wants telemetry.
What you have is a gray pattern. The same pattern that caused the EU to clarify its cookie law to require Yes and No choice to be equally accessible. And what you have is "Accept all" and "See details". Except that yours is worse by being a one-time notification.
To anybody complaining that this should be opt-in: opt-in telemetry doesn't work. The reason for this is that most people don't care, but they don't care either way. They don't disable it when prompted, nor would they enable it manually.
The idea of telemetry is being able to prioritize the work that will be most widely useful. For this you need a good and balanced sample of your users. You don't really get any kind of sensible sample if you only do it opt-in. Additionally, this ship has long sailed, everybody does opt-out.
What I do think however, is that it should very clearly notify the user of this, and give them an easy way to disable it. Like in OctoSQL[1] (disclaimer: which I'm the author of) which prompts you on first run and shows explicitly how to disable it.
All things considered, this is an open source project, so you're free to maintain a fork without telemetry. The Go toolchain also uses the Google-hosted module proxy by default, which really is a bit like telemetry already.
[0]: https://news.ycombinator.com/item?id=34707583
[1]: https://asciinema.org/a/eWQsyXQKi1fmithyTekAD5fWS