[urthen]

It performs an auto-update check. You know it's just performing an auto-update check. While there's not no argument to be made here, it's not like it's secretly collecting a bunch of data and reporting it before letting you opt out.

The paranoid, accusatory tone in the issue does nothing to help your argument, and I believe makes it harder for maintainers to take you seriously. Your argument is well thought out, and I don't necessarily disagree with it in principle, but you're making mountains out of molehills here. If you'd simply pointed out the issue and asked that the auto-update check be performed after the consent check - especially if it came with a PR implementing that - you'd go a lot further.

[sneak]

> You know it's just performing an auto-update check

Actually, I don't. I didn't agree to let it phone home. It does it before it tells me anything at all. It's also connecting to the telemetry service endpoint immediately on launch, despite not having selected a telemetry consent setting yet.

How about this one? It phones home when you click "don't phone home".

https://github.com/atom/atom/issues/20185

[urthen]

Don't use Atom then? Seems pretty obvious if you care that much.

I respect your privacy views and even share many/most of them (if not so strictly), and appreciate that you are probably just trying to bring this to people's attention. However, you are not the king of open source, and you cannot dictate how other authors must write their projects. It's an open source product written by a for-profit company. There's always going to be trade-offs.

[squeaky-clean]

Please don't make the "Don't use atom then" argument in response to someone sharing unexpected info about Atom. No one is saying that they are being forced to use Atom. They are sharing info others may not know about Atom.

You're saying "Don't use Atom", but what you're implying is "Don't talk about Atom"

[CKN23-ARIN]

> However, you are not the king of open source

I am, however, the king of my own computer, and this software does not respect my sovereignty.

[d0lph]

You are the king of your computer, but if you don't create rules for your subjects they will do as they please.

[dvfjsdhgfv]

That's very true. However, for several decades users had a bot more trust in FOSS as not betraying the user in the same way commercial apps do. Atom is abusing this trust. Just because Microsoft is doing what they want in Windows world it doesn't mean they can do the same when flirting with developers and preaching their love for FOSS.

[sneak]

Homebrew. Etcher. Gatsby. Atom. Syncthing.

Lots of programs are abusing user trust.

[urthen]

Like I said - if it matters to you, don't use it.

I've never had a problem with people that have extremely strict views of privacy. In fact I usually think they're right. But it's incredibly impractical, and most people don't currently care, which is the real problem. Treating relatively trivial things like this as if it were some conspiracy on behalf of the surveillance state doesn't do anything to get people to listen.

Get people to start dropping Facebook and Google, then we can start tackling the smaller offenses like this. Until then, treating this like you're going up against the NSA is a vanity exercise.

[catalogia]

> "Like I said - if it matters to you, don't use it."

Or he can do one better: Not use it and publicly criticize it.

[sjwright]

That’s not “better”, in fact it’s not really in the spirit of open source, where criticism comes in the form of patches and forks.

“Doing one better” would be patching and compiling from source yourself. Better still would be maintaining an up-to-date fork for the benefit of other like-minded people.

[sneak]

That's very neat, but it's not simply about me. It's about thousands and thousands of other, much less experienced people who use these software packages (Homebrew and Etcher and Atom in particular). They deserve privacy too.

I think most people don't realize. That's why these things do it silently.

Most people do care if you ask them. They don't want to be spied on. Given the option, most people will say "no, thank you". That's why these maintainers are so afraid of opt-in. They pay lip service to respecting user consent, but it's just that. They don't actually want to accurately reflect user consent, because they know that if they actually ask to measure it, they don't really have it. It's just like shitty web shops that automatically sign you up for their weekly marketing promo newsletter because you bought something one time. You don't want it, they know you don't want it, but they're still going to rob you of your time deleting them until you finally click Unsubscribe. Same deal.

Did you watch the John Oliver interview with Ed Snowden?[1] The show went out on the street and most people had no idea who he was or what he did or the information he released, but when asked if they knew that the government was logging all of their dick pics (which Ed gracefully confirmed that they are, in fact, doing), they said that they did not know and were not okay with it.

I think the core issue is ignorance, not apathy. I intend to educate people on the matter, and give them actionable steps to take to express their displeasure to the maintainers and their parent organizations. We can solve this issue before it gets any larger.

[1]: https://www.youtube.com/watch?v=XEVlyP4_11M

[saagarjha]

On the flip side, if you ask nicely I'm usually more than happy to give you some sort of usage information if it seems reasonable. If you don't, you sure aren't going to get anything from me, and I will go out of my way to stop you from collecting that information and tell other people about what you're doing, because you have clearly shown that you cannot responsibly collect usage data.

[sneak]

> However, you are not the king of open source, and you cannot dictate how other authors must write their projects

No, but I can convince other people who don't like their computers being used a spying tools against them to put social pressure on maintainers so that they stop doing this nonsense.

Atom's telemetry used to be on by default; spying silently caused them such a shitstorm that they added a consent dialog. They're almost there! Now they just have to make it functional.

https://github.com/atom/atom/pull/12281

There are precedents. We can push back, especially against open source projects.

> There's always going to be trade-offs.

I don't think that's how software works. I certainly don't take that to mean that I should just accept that it's going to spy on me. I don't want that, and I don't accept that, and I will yell, loudly, at anyone who says I should accept that without a fight.

[sjwright]

It’s open source. They aren’t forcing you to use their software, and they’re certainly not forcing you to use their binaries.

[mirimir]

But here's the thing: They don't disclose that they'll check for updates before asking for permission.

So how would a user, who hadn't seen this discussion, know that their "we'll ask" reassurance was bullshit?

[sjwright]

If you actually care about that—as opposed to complaining for the sake of it—you shouldn’t be allowing new software to establish outgoing connections without your consent anyway.

So if you actually did care, you the user would know because your operating system alerted you.

Nagging everyone to comply with your interpretation of “the right thing” might feel good to you, but it’s actually a very weak and porous form of security theatre.

[sneak]

i use Little Snitch and DNS blackholing of tracking services, so these issues don't really affect me.

today i decided i care more about strangers than i thought i did

people brand new to our industry should be able to download a nice gui editor and not get their consent trampled and get spied on even after they click the "dont spy on me" button

it's easy to laugh and be like "oh lol it's software from microsoft what did they expect, noobs" like someone else did in this thread

but that's bullshit and you and i both know it

the software simply shouldn't do that when you say "don't send my data away pls"

i don't want everyone to have to say "oh use homebrew it's great but also add this weird line about analytics to your .bashrc before you install it oh wait you don't know what a bashrc is huh" when they talk to some teenager who just got a $15 rtlsdr and wants to install gnuradio on their mac

that's not a good first-10-minutes-at-the-command-line experience.

i don't think that's fair or good or optimal.

i want the world to be different, and i want these maintainers to realize that they made a mistake, and revert it. i don't think they're bad people, i think they're just misguided, and they're optimizing for vanity metrics like user count, which will effectively go away entirely if i succeed and they only get telemetry from users who said "yes it's ok i don't mind". that's a lot fewer users, and they know it, which is why so many of them are refusing to engage with the ethical argument about silently using a user's own hardware to spy on them without their knowledge or consent.

it shouldn't be a controversial position that our tools should not spy on us.

[bromuro]

Thanks for defending your points.

I don’t understand why some here take your reports on GH personally. The fact that your reports aren’t taken seriously by the Atom team worries me. Your battle is right and presented in the right tones.

[BubRoss]

> i use Little Snitch and DNS blackholing of tracking services

Why do you need that when you can just not use things that spy on you (by your own logic)

[BubRoss]

This is about not lying to your users

[sjwright]

It's not a lie if they aren't logging anything.

(I'm not saying that's the case, by the way. It's just an observation. We can't really know if they're lying or not—and to that extent I agree with your point.)

[taneq]

How are you meant to know about this a priori unless you've read an article about it on an obscure tech website?

[edoceo]

Use strace on every process? Inspect all code that ever runs on your boxen? Those work but, I still prefer the canary-article. Much easier.

[taneq]

What'd be really nice is transparent, built-in sandboxing of every non-system application by the OS. The desktop security model of "protect your files from other users but not from applications you run" is horribly outdated.

[matheusmoreira]

On top of all that, we also need deep packet inspection and filtering. All the data flowing into and out of the sandbox must be inspected while in the clear by filtering software under our control. If the packet is known to contain nothing but a unique identifier for tracking, it gets blocked. If it also contains useful data, the identifier is either deleted or anonymized before the packet goes through.

ISPs, companies and governments can do it for surveillance, censorship and security reasons. We should be able to do it too in order to empower ourselves.

[rossmohax]

flatpack on Linux has all ingredients to enforce such sandboxing

[marcinzm]

>However, you are not the king of open source, and you cannot dictate how other authors must write their projects.

And you're not the king of what people are allowed to voice their views and opinions on in public forums.

Authors can do whatever they want and users can say whatever they want as a result. Doing something does not mean you are free of the consequences, including people voicing their opinions, of those actions.

[hashkb]

The maintainers who reply dodge the issue in a way that must be frustrating to the reporter; and makes it seems like they'd probably not accept a PR that modified this behavior. Calling privacy advocates "paranoid" in today's climate is a bit suspect. There's a definite issue here that the Atom team should address - a reasonable user would expect that, after opting out, the app would never phone home, and it does.

Blaming tone is too easy - at this point the Atom team is representing Microsoft, so I'd say the burden is on them to soak up a little snark; especially coming from a user who maybe expects them to behave a bit more like the GitHub of old. Even if they were a small open-source team I would still expect them to directly confront the issue instead of beating around the bush. It's about privacy, and splitting hairs to deny the reporter's reality is a bad look.

[matheusmoreira]

> Calling privacy advocates "paranoid" in today's climate is a bit suspect.

It's an insult. They're comparing people with legitimate concerns backed by evidence and precedent to paranoid schizophrenics. They are implicitly saying that people who value privacy are delusional and mentally ill.

[urthen]

The application clearly states that it sends an opt-out notification anonymously after opt-out. I think they're doing better than most in terms of transparency. Is there room for improvement? Sure. But I don't think attacking people who are already doing pretty well compared to the field is the best use of anyone's time.

This is just picking on an open-source team which already has shown themselves to take user feedback about privacy seriously. If you are so suspect of third parties that any device or application sending any network traffic without previous authorization is of serious concern to you, there's a lot more important and impactful products to voice concern about. Like pretty much any other part of Microsoft.

[lukasb]

The "never phone home" UI should make it clear that it will still phone home to check for updates.

But I'm sympathetic to Github here. Having users on unpatched software is a bigger risk to them than not having 100% perfect insulation from sending their IP address to Microsoft.

[vunie]

>Having users on unpatched software is a bigger risk to them than not having 100% perfect insulation from sending their IP address to Microsoft.

That's not a decision for Github to make. Let the user decide. Perhaps by even presenting them with a dialog on first run that informs them then asks them ti decide.

[Multicomp]

I agree wholeheartedly.

Somewhere out there (prog21 ?) is an article stating that the one in desperate need of electricity (the computing device) is subject to the one in a position to provide it (the user), and I agree with that sentiment.

I'd even like an IETF-standard-like T-shirt that says 'The user's will MUST be obeyed as faithfully as possible unless prevented by unrecoverable circumstances' or something like that.

Click-through EULAs and dialog boxes are another symptom of the 'elite-developer-itis' our industry can sometimes exude.

[sneak]

Users having unpatched software is no risk to GitHub. It's risk to those users.

[TAForObvReasons]

The more interesting question, which you completely sidestep by trying to categorize the leaked info, is whether software should perform an auto-update check before giving you an option to opt out. I don't think the author would have been concerned if they were asked whether the software should check for updates.

[bromuro]

MacOS apps are used to prompt a dialog - atom should too.

[egdod]

It’s reporting your IP address. That’s not nothing.

[na85]

What's the alternative? I'm not sure how a client can query a server for updates without exposing its IP address beyond obfuscation like TOR or perhaps a VPN.

[hashkb]

It can just obtain the user's consent; and, if denied, the alternative is to not phone home at all.

[mirimir]

That's pretty much the norm in Linux.

One typically gets packages from the distro's repository. And traffic with modern repositories uses HTTPS. So third parties don't see what packages are being used. And unless one configures a developer repository, there's never any traffic to the developer.

Also, IP addresses are considered PII under GDPR.

Edit: Still, if one cares that much about ones IP address, it's prudent to use a VPN service or Tor.

[jdnenej]

IP addresses are only PII if attached to other information. I can generate a list of IP addresses and store/sell that data and it's all fine because its meaningless without associated data.

[mirimir]

The other information is "using Atom".

[colejohnson66]

> Also, IP addresses are considered PII under GDPR.

Not true. They’re only PII if they can be used to connect something to a person. It’s a minor distinction, but it is one nonetheless.

[jdnenej]

In this case it would not be PII data if it was just an IP address in a webserver log saying someone checked for an update. It would be PII data if it was linked to your GitHub account.

[serf]

an alternative updating mechanism that doesn't directly expose IP of the updater?

a dht/blockchain mashup comes to mind, but really it just pushes the IP knowledge to knwoledge of some other intrinsic variable that can be used for fingerprinting.

the issuer of public information (an update) doesn't necessarily need to know who collects the information, it's just architected like that for most internet things.

[gen3]

That’s just not feasible for most projects. Why ship a blockchain along with your product, when you could just go with a GET request? The technical complexity, server space, and additional bloat that would be needed to be added to the app would be a waste, IMO.

[egdod]

>”Would you like to check for updates?”

[sjwright]

Or you could maintain the current version number on the Wikipedia page for your app. Your app could then request its wiki page over HTTPS as a check for updates.

Then the only useful logs would be held by Wikipedia. And more broadly, requests to that page would be utterly lost in the noise.

[dane-pgp]

That's not a bad idea, actually.

Ideally the article would contain (in a comment or some hidden field) a signature from the PGP key of the app's developer, covering both the version number and the date of release (to stop replay/rollback attacks).

Alternatively, the app could look at the article history and find the latest edit made by the developer's Wikipedia account, so that malicious edits would be ignored. This assumes the threat model doesn't include rogue Wikipedia admins rewriting history or hijacking accounts.

At the expense of a smaller anonymity set, it might make more sense for the app to query Wikidata instead of Wikipedia: https://www.wikidata.org/wiki/Q16766305

[sjwright]

> (to stop replay/rollback attacks)

In my imagining, the content of the Wikipedia article does nothing more than trigger a notification to the user; it would be the user's choice whether to initiate a network connection with the vendor's server for the "real" check and binary download.