Debian's Chromium changes default search engine to DDG
mail-archive.com
mail-archive.com
> This change caused my chromium browser to report that it's being managed by my "organization". I thought that my machine was somehow compromised. This is terrifying! I wound up deleting my entire chromium profile before I discovered that the root cause was this DuckDuckGo config change.
Tell me what organization. "Debian" would have been perfectly fine to show here.
I really don't understand why Debian wouldn't just patch the source code to change the default search engine rather than shipping an enterprise policy like this.
You can see the details in the chrome://policy page.
I should've thought to post a Debian bug report after that (especially since the Debian bugs database was one of the first things I checked). I'd reported the cause informally to some colleagues, and then must've gotten distracted with what I was trying to do before I saw the suspicious message.
Chromium does change the message to "This device is managed by <google workspaces domain>" if attached to Chrome Enterprise.
Moreover, defaulting to Google is also a policy.
EDIT: Apparently the Debian team agrees: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=956012#72
Just because Debian thinks that they are doing the right thing, they are in fact, controlling the user's settings through a policy It's just telling you that, as it should.
The notion that it has anything to do with keeping search the default or not is like, such a silly assumption i don't know where to begin.
The feature overall came from a desire of enterprises to manage browser settings. Back then, Google was one of the first to tell you someone was doing that to you so that you knew your organization could see and control your settings.
IE had a deployment kit that let you deploy managed browser settings, but you didn't get told (this changed, eventually, i think, it's been a while)
Letting orgs change the default search engine was an explicit, designed goal, since some wanted to redirect people to their internal searches by default, etc.
There is in fact, another way to do this that is easy and doesn't give the user the same warning, and is meant for software distributors
You can just use master preferences here for this kind of thing and it is meant for this use case.
Google in fact, made this easy and officially supported, despite your claim.
It would likely be pretty silly to make this hard - end users aren't using these interfaces or tools, and distributors always know how to change this stuff .
As HN as grown in popularity, the sheer number of kneejerk reaction comments has unfortunately kept pace (IE the overall percent has not dropped. Even sadder, nobody ever goes back and edits it or replies and was like "you know what, i was probably wrong".
They feel comfortable moving on and doing it again.
ah, so it should also be displayed if it's set to Google's preferences
they're "someone else" as equally as Debian
So Google's defaults are "un-managed" and imply no control over users' settings?
Why shouldn't the Google built Chrome binaries display the same warning and a binary from a specific linux distro should?
It says "Managed by your organization." every time you click the menu button. It's even in a different background color and bigger margins.
If you click it, it says "Your administrator can change your browser setup remotely. Activity on this device may also be managed outside of Chromium. Learn more"
I made the policy myself and copied it myself. The remote part is purely conjecture on part of Google. The second part about further management also feels out of place.
Then on every setting that's been changed by policy there's a building icon that says "This setting has been changed by your administrator." That's probably the only message that needs to be there.
On its own it seems innocent and useful, but compare this to the number of steps you have to take to eg.: delete Google cookies.
Accurate conjecture nonetheless. If a 3rd party has dropped a managed policy file onto your install, it's reasonable to assume they can drop an updated one which does all the scary stuff at any time. It's a waste of effort for Chrome to parse the current profile when the mere presence of one can compromise security.
Nothing to see here; we can move along.
But by default isn't Google controlling the user's setting through a policy. The policy includes, among other things, the rule that Google is the default search engine. (Needless to say, that setting is a cash cow.)
And in fact, the user that manually changes the default policy herself, i.e., changes the default search engine to something other than Google, versus delegating the decision to a third party such as Debian or Google, cannot remove Google as a search engine choice! It is greyed out. Permanent. Chrome binary allows users to delete all the other choices, except Google.
If you read the thread linked in the comment you replied to [1], the Debian folks added a JSON file to `/etc/chromium/policies/managed` when they should have edited `/etc/chromium/master_preferences` (a different JSON file). If they had done that, there would be no message.
Everyone on the thread agrees that's a better solution.
I love the project, I really dislike this 'gather things that might be useful to someone' behavior and having to MITM my system to see what it's actually doing
It's superficial but I completely understand the alarm
For those noticing this, be aware that you'll likely see it elsewhere. Don't panic.
They don't enforce organization policies, but they set the default config to "no search" and then leave it up to the user to change it.
And you are assuming there is only 1 user in the machine. Which is not the case for people who have kids.
It worked great for cookies in the EU. Really improved my browsing experience.
There is a spectrum between everyone votes on everything and yielding everything to our corporate overlords. For example there could be an "I don't care" button which picks one at random.
Cookie banners despite their annoyance are educating people and motivating corps to minimize their PII usage.
I personally improved my browsing experience with the Consent-o-matic addon, which knows some of these cookie modals, and fills them for me according to my preset (which is to deny all cookies).
Also, people not caring should not be a baseline. People are ignorant about the vast majority of things going on, and yet they are much better off if those are taken care of by people who actually give a damn - as evidenced by a good deal of regulations that keep lead out of paint, antibiotics out of chicken, and so on.
No disagreement there, I just think the GDPR rules that led to ubiquitous cookie consent dialogs were indicative of policymaking by people who didn't fit that category. It should have been easy to predict this outcome from that policy.
They haven't been caught running fingerprinting scripts yet and they dont have an account system to tie to your searches. At best they could use your ip to build a shadow profile and thats wildly inaccurate in our mostly ipv4 world.
And I disagree with your premise that it's particularly difficult to link a persons IP to their real world identity. There are organized fraud gangs who have it down to a science. know exactly what dept. of the ISP to call, what to say, etc. Basically if someone knows your IP and your ISP account is registered in your name it's game over.
I just want to avoid my data being monetized.
uMatrix shows a 3rd party request to improving.duckduckgo.com every time I visit a page from DDG search results, ostensibly to measure click-through rate. This is claimed to be anonymous, but in principle it gives DDG the opportunity to log much about their users' browsing habits.
It is still, I would claim, objectively more private to use DuckDuckGo than Google even in a world where they are lying about their privacy policies, purely because DuckDuckGo does not have the same surveillance scope and level of infrastructure as Google.
And that's really what we're arguing about here, unless you have a more private alternative to DuckDuckGo that has been subject to more rigorous audits and can scale to support being the default search engine for a bunch of nontechnical users?
What audit would you suggest for us to prove that statement? It would be great to have a some independent party verify this.
Cynically speaking, I am not sure that there is an audit you're going to be able to do that won't cost a ton of money that the people in this thread would trust as definitive "proof" of anything[0].
I think a big part of what I'm personally getting at with the comment above is that I'm not looking for perfect proof of anything; independent audits are great and I love to see them and I absolutely encourage them, but remember that the point of comparison here is Google/Bing. Take it with a grain of salt, and purely opinion me, but I think its fine for private search engines to offer the best proof of their claims that they can and to otherwise ignore people who demand perfection or nothing.
It's great to see more search engines in the space with a focus on privacy, and if you're able to pull off building your own indexes, that's also a pretty big win. I wish there was a more obvious path forward for your company to make money (I get nervous when companies say, "we'll figure out funding later", to me that comes across as a little bit of a time bomb). But in general, always good to see more private options for people available.
If I was in your position and I was looking for audits, I'd honestly be looking at the same sources that DuckDuckGo's founder talks about further up-thread, because that would at least allow me to say, "the same sources that claim DuckDuckGo is private have also said that we are private." But it's not my area of expertise, so maybe that's bad advice.
I’m guessing they’re intentionally degrading the non-Javascript page as an anti-bot measure, but it’s so bad that I find it disingenuous to suggest that the non-Javascript page even a valid alternative at this point.
edit: I should have just down-voted and moved on.
I mean - yea. Maybe your dentist never actually graduated dental school. Did you call the dental association to check he’s a member? Anyone can just print out a certificate on their home printer and put it on their wall, y’know. And even if he is, do you think the dental association actually called his college to verify the transcript he gave them when he joined in 1988 or whatever?
You really should do an independent audit of your dentist’s dentistry skills. Perhaps you should demand he does some kind of standardised test. But he can’t just go to a testing centre, you have no way of verifying that. He must do the test in front of you.
And how do you know the answers to the standardised test are correct, anyway? You will need to do a dentistry degree yourself first.
TLDR: A trustless society doesn’t work, and most people aren’t out to pull one over you.
Just to be clear, are you saying that given the choice between collecting data and lying about it vs. collecting data and being explicit about it, you’d choose the first option?
Yes. Absolutely. Because that would give me some legal recourse.
Would you hire someone who hides in the fine print they can steal from you and you can't do anything about it, or hire someone else and accept the chance that they might steal.
The choice is between a bad thing definitely happening, or a bad thing possibly happening.
Also a lot of what we do is open source on GitHub. We recently put out a help page detailing or web tracking protections that link to a lot of the relevant repositories: https://help.duckduckgo.com/duckduckgo-help-pages/privacy/we...
And finally, I’m not sure that random or just popped up is an accurate characterization for us. We’re pretty well established at this point, having been around for nearly 15 years! I was an early user of this site and a frequent contributor during the early days of DuckDuckGo.
The best thing you could do, if you actually care about privacy and not just $$$, is to open-source the entire search index db and accompanying webserver software, making it easy for users to setup their own local instance of DDG which is actually auditable. Additionally, posting a notice on-site which notifies your users that their searches may be recorded and tracked in spite of what the privacy policy says(due to the USA jurisdiction of the company making it susceptible to National Security Letters and secret gag orders) would be the right thing to do.
> The best thing you could do, if you actually care about privacy and not just $$$, is to open-source the entire search index db and accompanying webserver software, making it easy for users to setup their own local instance of DDG which is truly auditable.
self hosting isn't feasible for 99% of the population. DDG is aiming to be the mainstream privacy protecting search engine, I used them for a while and can appreciate their efforts. if you want something nerdy and and self hosted use a searX instance or host it yourself.
Its only this way because companies have a vested interest in keeping it like that. It's how they make their money. It is absolutely within the realm of possibility that people host their own search engine. 99% of people know how to install Google Chrome right? this should be no different. The entire search engine & webserver stack it depends on could be bundled into a .exe/.app installer with simple instructions people can understand. Consider XAMPP- which already provides a webserver stack that is extremely easy to install on Windows/Mac just by a simple .exe/.app that 'just works'. This hypothetical search engine could use similar methods as the XAMPP installer. There is no technical reason why this can't happen. It just isn't happening because it'd increase competition, cutting into DDG's profits.
Everyone has a PB of SSD disk space, some few TB of RAM and a few thousand CPUs to throw at the search problem, or is happy to type in a search query and give a 16 core CPU a few days to execute it, right?
That is just a naive implementation. For the first 10 results you grab ads, the database of those is significantly smaller, for the next 20 results you look at Wikipedia and stackexchange clone sites. Everything after that is indexed using math.random(). If you want to get fancy run the query through a fact creating AI and present the results inline, people are always happy to know that the color of the sky is purple or that the ideal amount of chess players is 5. Disclaimer: I have never seen googles source code nor any patents related to it, any similarity with existing search engines is pure coincidence.
Easy to self-host? How large do you suppose the Bing index is, for example? Simply storing the index would be an immense undertaking beyond the reach of probably everyone who has ever self-hosted anything, ever. This ignores the compute required to actually search it, as well as how it would get updated.
I'm not sure your request is remotely reasonable.
Secondly from 2000 - 2018 the internet went form having ~17.000.000 unique domains to having ~1.600.000.000 unique domains. see: https://www.internetlivestats.com/total-number-of-websites/
The performance for desktop computers have actually not increased as much as you would think: https://www.karlrupp.net/2015/06/40-years-of-microprocessor-...
Your assumption is correct if you look at supercomputers, where the fastest in the world in 1999 could produce ~2.3 TFLOPS and in 2018 it could produce 122 PFLOPS which is around 5000 times the increase in FLOPS.
But i doubt most of the people you would want to go through this index has access to a super computer.
- they are less likely to throw a captcha in your face if you connect over VPN
- they have less surveillance infrastructure and run less code clientside than Google does
- they are at least not explicitly tracking you
- they have a lower number of secondary data-points from other services that can be connected to your searches
the list kind of goes on. I don't assume that DuckDuckGo is perfectly trustworthy just because they say so, but Debian has a choice of a couple of different default search engines that are mature enough and give good enough results to use as a default search tool: Google, Bing, DuckDuckGo, etc...
Of those choices, DuckDuckGo seems to be a pretty reasonable decision.
At the very least, DuckDuckGo lets me search when I'm behind a VPN and have anti-fingerprinting tools turned on, Google very often doesn't. It's not a super-hard decision for me which one is more private.
Do you want to try and argue that Google is a more private search engine than DDG outside of the US?
I am not claiming that DDG is bad or anything, I just don't like putting trust into something just because it says "you can totally trust me!".
What would we have Debian change the default search to? I get that DuckDuckGo might not be ideal, but it is better than Google, Bing, Yahoo or Marginalia. The results need to be good enough, but also not obviously anti-privacy. It basically leave you with DuckDuckGo, Qwants or Ecosia. Personally I might had picked Ecosia, had they not had a cookie banner.
Unfortunately my search engine is far away, both in terms of functionality and hardware capacity from being able to deal with that. Maybe some day, who knows, but not yet. Even if I'm destined to make the Linux of search engines, we're metaphorically living in 1992 or so.
Would be funny though because it's both developed on a Debian workstation and hosted on a Debian server.
Do you have any stance there? (I'm not saying you should have one or agree with mine, just curious. Every search engine might have its time and place).
That said I do block some sites, mostly nazi stuff if it's designed in such a way that it crops up in regular searches. It's a fairly small number of sites though.
You can check if a site is blacklisted using a site:-search like
I see this sentiment a lot lately and it seems weird to me that since narratives don’t need to have an inch a truth to be constructed. Narratives can be fabricated from made up atoms if need be . If crackpots wants to believe they’re being prosecuted , they’ll do it anyway. The ammunition you fear you’ll give them is one they already possess if they’re already too far gone.
I worry about the sustainability of such a service though. Don’t they inevitably get blocked upstream?
It might work better for more local organizations to host such projects. I’ve always liked the idea of community centers and churches and whatnot hosting shared services for their community.
What’s the problem? What matters is that my searches aren’t recorded and added to a profile.
But yes, I do use !g when needed to jump to Google. It's just the default that works better on DDG for me due to the Incognito window issue - so much so that I've been running DDG by default for quite a long time now, just for that reason.
For me this is terrible. I can scan through 100 search results pretty quick, going 4 pages deep is something I have always done often. Seeing the reported ~82 million results shrink suddenly to 4 pages and maybe 389 results with ommitted included is extremely disheartening. But at least google scholar still works properly.
When I found out about this I was so pissed I made a crappy little website to more easily complain about it to people. http://googlesearchonlyreturns400results.lol/
> Hey let’s change the default engine to DDG
> I’ve used DDG for a week, let’s do it!
As a CentOS user, the Ubuntu snap is updated much more often than the EPEL package.
Anyway, so now I just manage my own updates via the tar.gz because I can't be bothered to rebuild .deb for the releases and hope snap and its flatpak friend go the way of the dodo
However, I would prefer the default to be "no search", and to only search if the user explicitly specifies which search engine to use. (This does not necessarily mean that Debian has to do this; it only means that it is what would be my own preference. Some other people will agree with me, although some people will disagree.)
Regardless of the default settings though (sometimes different default settings might be suitable due to the distribution; in this case it doesn't matter, but for some settings of some programs, it will matter), the end user should have the opportunity to change all of the settings.
I don't have links stored in history but it was fairly recent, it was their CEO on twitter I believe and they got a LOT of backlash.
In hope to quickly clear up some common misconceptions about them though: we don’t censor, we don’t move things so far down that they are effectively censored, we don’t have any definition of misinformation, and we don’t rank based on any political agenda or opinions (that includes mine!). This is just a summary though so would read the help page for details.
Now today since I've linked this thread, you have now changed who could reply.
It's pretty clear that you will say or do anything to try to save your tanking company.
Also, all I'm doing in that old thread (and this one) is doing my best to explain how our news rankings actually work. Again, currently the best and most thorough explanation is in this help page (more than can easily be put on Twitter): https://help.duckduckgo.com/duckduckgo-help-pages/results/ne...
I'm sure many people have thought about it, so, I wonder, what is the issue with this approach?
There are a few people (like me) who are crazy enough to give Linux desktop systems to real end users. Although I suppose I could just set their search engines for them, but that's just a pain.
You’re not happy with the default which you can trivially change but no one has to change, so you propose having no default so you still have to change it but now everyone has to change it.
It's not unheard of though. See: Ubuntu.
Lots of folks misremember how the Unity Home Lens feature actually worked, and that the Ubuntu installer explicitly asked you whether you wanted it enabled.
If enabled, there was no presumption of local-only search. It was front and center in your search results, and the affiliate revenue was only collected when you clicked on that Amazon result.
Canonical claimed that the search was forwarded to them, and then anonymized to Amazon. That's more than what Microsoft seems to do with Bing results in Windows search, which you can't seem to disable without installing third-party scripts.
Was it a perfect solution? No - But it was a way to make revenue in an otherwise open-source project and free of cost. Its interesting how on some days, people argue for more money towards FOSS devs, and then others argue that revenue-building systems shouldn't be implemented. Yeah, Canonical makes enterprise investments, but they're obviously weren't in the realm of Red Hat in terms of enterprise support.
That said, it was a braindead idea to begin with. Not one person in existence wants or expects shopping links when searching their computer via a dock like interface. This is meant for applications or files.
When shopping, most are going to research on Google...or perhaps search a category on Amazon itself before choosing a product.
I'd be curious it Ubuntu even made a single sale from that venture.
I saw it as a way for Canonical to start bringing end-user-oriented features to the Linux desktop, something that Linux sorely needed circa 00s and early 10s. They also had the forethought to bring web apps to the desktop before most major players did.
It was all part of their 'continutity' push at the time: desktop and phone in one cohesive interface. Too early for its time, but its clearly working for Apple.
Yes but not for the reason you are imagining. I can think of a few things in common between everyone I've ever known to use Linux for personal use and general consumerism to improve quality of life is not one of them.
Prior art is that Linux Mint at one point had Yahoo as the default search engine in their Firefox builds. I am not sure whether that is currently the case.
It is opinionated choice, not a bug fix, and definitely not a security update (severity: wishlist). I don't have a problem with maintainers having opinions, but on a stable distribution like Debian, I would have preferred they expressed them before the freeze.
Users can always switch back to Google if they so desire.
It is literally against anything Debian has done the past decades
Why would they choose a proprietary software and proprietary service?
Chromium at least has a BSD like license and is open source [1], the search provider alternative should be similar
Literally nobody can confirm DDG's privacy claims, they refused every independent audits and your search is leaking to their servers [2], it happened again [3]
And let's not forget the Microsoft Bing trackers ;)
[1] - https://chromium.googlesource.com/chromium/src/+/HEAD/LICENS...
Or was that intentional?
That's a shame.
This should be predictable on the basis that it is a free service, making you the product, and somebody else, therefore, the customer.
It is hard to know what else one can do to get useful search functionality. It has been a long time since Google dropped any emphasis on usefulness. Any useful results seem purely luck nowadays. You cannot even buy a subscription to "useful" from them or from, AFAIHF, anybody else.
Doesn't this line of logic lead to every free site being evil, including the one we're talking on?
DDG has ads based on your search query, that's how they make money. The difference is that they don't profile you, at least that's what they say. You can spam and respect people privacy, just by not looking who you are spamming.
And yes, "privacy" is a marketing ploy for anyone who is not Google. As for general purpose search engines, there are only two: Google and Bing, most others (including DDG) are just a front for Bing. There are other, more specialized crawlers including Marginalia whose author often posts on HN, and there is Yandex for Russia and Baidu for China, but the general idea is that if it is not Google, it is Bing.
Many companies put out free software to drive people towards their products, and Chrome with Google Search seems to be one of those. As many know, improving and maintaining Chrome is not free, and having Google Search being a default is one part of what helps pay for this work.
Yes, this is Google, yes, this is likely a tiny drop in the bucket for them, but at the same time, it's taking away potential revenue from Google.
If this was some smaller company that produced a product that had some default that pointed to one of their SaaS offerings or the like, there would be potential issues raised over the Debian maintainers changing this default.
Well, thankfully this is Google, and not a small company then.
Google still has 92.5% global market share, 10x more than all other search engines combined https://radar.cloudflare.com/notebooks/searchengines-2022-q1
1. Within existing legal boundaries, of course
My take from a business perspective is that Google produces Chrome and Chromium for a number of reasons. Good will to the community (with how permissive they are with the license), and having a stable platform to be able to build things like GMail and Search on-top of. But there is also the Ads side that benefits from Google Search being the default.
So I guess there are really many benefits for Chrome's existence, and Google Searching being a default is only part of that. But I still stand by my original post and reasoning.
so complaining that it is being modified to restore it back to its demonitised form reeks of entitlement
many of us see this as a positive
Wow. Yes, that's how competition works. Are people at the tech monopolists really that entitled that they consider the entire world's purse strings theirs to control?
Good. Hopefully we will see anti-trust enforcement force Chrome and Android to not default to Google search either.