Please Google, understand this; I don't want my browser linked with my Google account.
Google Chrome now has a "feature" that will automatically reset your homepage, search settings, wipe all extensions and wipe all extension data whenever it detects an unexpected change to any of your settings files or they do not match up to the PC they are locked to.
The goal was to prevent third parties from hijacking your search engine within Chrome. Things like bundleware installers on Windows that switch it.
The actual result is that all your settings are now locked to a specific PC and will be reset anytime anything changes. As the developer of Google Chrome Portable, this means only your bookmarks move with you as you move PCs unless you're logged in. This also means that even if you backup your Chrome install, if your PC dies, you can not restore that Chrome install on another PC. Your bookmarks will be restored, but you will lose your settings, extensions, and passwords since your profile was moved to a new PC.
This stands in stark contrast to Firefox which is completely self-contained. If a third party bundleware installer hijacks your search engine or homepage, you can simply set them back as they should be. All your extensions, extension settings, etc will continue working as expected whether or not you opt to use Firefox Sync.
You absolutely must link your Google Chrome install to
your Google ID on Windows or you will lose data.
Incognito mode at all times, no Google ID ever, trash local history on every exit. I don't carry the expectation that my bookmarks must live on and follow me from machine to machine.
I don't care about what my search preferences are. I simply turn address bar searches off, or point it to localhost. I don't conduct rampant web searches from the address bar, because an address bar is an address bar, and I only feed it well-formed URLs. I don't waste my time regularly using web sites that can't provide me with a URL that isn't intended to be typed, and sub pages that I can't readily navigate to.
Address to long? No soup for you. Absurd URL parameters? No soup for you. Deeply nested pages, locked away behind a login, 30 clicks deep, and obscured by dark patterns? No. Soup. For. You.
Preferences are preferences, and that's all they are. The default install should be completely useful and practical as-is, on a clean install. If it isn't then preferences should be exportable to plaint-text using an open standard.
Firefox happily dumps its bookmarks as simple JSON. As a result, when I want to transport my bookmarks I can, although, digging back into my archive more than a year 50% tend to perish due to simple link rot.
I think web applications should never have existed. The web was designed for hypertext, images, and other static consumable content. It wasn't designed to be a platform for massive distributed applications. Why are there so many security fuckups on the web? Because it wasn't designed to be secure. The fact that people are now programming desktop applications in JavaScript against NodeJS suggests to me that the web has gone too far. We should have built something new to handle what we now call "web applications".
I, too, look at huge disgusting URLs every day and wonder why we didn't see this coming.
Being that I'd never intended to 'sign into Chrome' in the first place (the form looks awfully similar to the gmail login), I essentially rage-quit Chrome and haven't looked back since. I've embraced Firefox and I feel better using free and open source software.
Performance has been excellent. The only thing I regularly notice is that many startups seem to only QA their products in Chrome.
They've actually gotten very upfront about their user features in recent versions. It's a great toolset for maintaining different sessions with multiple sites. Or for using as an incognito session with less ephemeral data.
Additionally, I've found FireFox to be more consistent about it's quality, vs having occasional releases with a jump in bugs/instability.
I have multiple g-mail accounts. One of them is my primary personal one, and another one I use is a shared account that my band uses for business and soundcloud.
Google makes it hellish to sign out of one and into the other. For the longest time, I had them linked, and my band mates could see my personal e-mail. I don't know if they could get past a log in point, but the fact that its so hard to keep them un-linked and log into one or the other is infuriating.
You can change users by clicking the top right of a chrome window.
They may not be obvious to set up, but they are worth learning about if you are going to use stick with Chrome and use multiple Google accounts, or multiple accounts on any service - Twitter, FB, etc.
The total market share of people who care about whether or not their browser is linked with their Google account is probably less than one percent. The value they get from knowing everything you do online? I'm not sure how you'd calculate that, but I'm positive it's greater than losing the tiny number of hackers who are protective of their personal data.
Google will continue to build the browser under the assumption that your browser will be linked with your Google account. If you don't link, you'll be a permanent edge case.
If you don't want to link your account, it's much easier to just switch browsers.
> The value they get from knowing everything you do online?
This is a myth. Or less politely, a lie you have been told.
https://www.google.co.uk/chrome/browser/privacy/ is very clear about exactly when chrome reports back data, including what happens when you sign in. It does not report your online activity in ways different from other web browsers. It sends just about the minimum possible to implement the features it provides.
I don't claim to know anything about how Chrome works, so could be completely wrong in this, but my sense is that it is part of a broader desire to tie together tracking across all devices. Cracking that nut is very difficult from an analytics perspective, and having people logged in to Chrome across all devices provides a very convenient GUID for matching against AdSense units, DFA tags, GA tags, etc.
So the value in this play is not necessarily in knowing everything you do online, but simply knowing that you are who you are regardless of device or Chrome instance. That means data about you is consistent, and can feed into the broader targeting algorithms to serve you the right ad, at the right place, at the right time.
Again, I don't know what data Chrome makes available to Google's ad services, so please tell me if I am in fact incorrect here.
If you read the very detailed Chrome privacy policy (https://www.google.co.uk/chrome/browser/privacy/), you will see that it does nothing different to other browsers here (each server sees your IP address, cookies, etc). It does not send any form of uniquely identifiable ID with every request. Only some very specific cases (offline drive, google wallet, DRM auth, etc) send your signed-in userid, which they need in order to work at all.
Please read the privacy policy. People put a lot of time and effort into writing it and it answers all your questions.
I see a section in the link you provided that reads:
> "Chrome OS may send a non-unique promotional tag to Google periodically (including during initial setup) and when performing searches with Google."
Sure Chrome may not be sending a GUID with every request. That wasn't my point. To clarify, my point was that Google has a GUID (your Google account, and likely other forms of ID) that identify you pretty clearly across devices/Chrome instances. I saw nothing in the policy you linked that said Google AdSense and other Google ad services (DoubleClick tags, GA, etc.) do not use that information to link identities across devices for purposes of tracking.
The AdWords and GA teams have rolled out significant advances in cross-device tracking over the past year or two, including new metrics and reports in the respective platforms. It seems pretty obvious to me that this is a data point that is factored into making that link.
But again, I'd love to be proven wrong if I missed a section of your link that explicitly says otherwise.
> To clarify, my point was that Google has a GUID (your Google account, and likely other forms of ID) that identify you pretty clearly across devices/Chrome instances.
I think the privacy policy is quite clear that Chrome does nothing here that other browsers don't do. Obviously if you have a Google cookie from signing in to the website then that's going to identify you pretty clearly, but that's just how cookies work, and would be exactly the same in any other browser. I'd like to specifically direct your attention to the first three paragraphs of the policy, and in particular:
"If you use Chrome to access other Google services, such as using the search engine on the Google homepage or checking Gmail, the fact that you are using Chrome does not mean that Google will receive any special or additional personally identifiable information about you."
> I saw nothing in the policy you linked that said Google AdSense and other Google ad services (DoubleClick tags, GA, etc.) do not use that information to link identities across devices for purposes of tracking.
I wouldn't expect to find any information about how AdSense works in the Chrome policy. The privacy policy for ads is over here and discusses this topic: http://www.google.co.uk/policies/technologies/ads/
https://addons.opera.com/en/extensions/details/download-chro...
It's called Firefox.
Also, chromium seems to be getting less and less care from google now. See this bug for example which prevents using screenshare in hangouts: https://crbug.com/416856
check this comparison between chrome and iron (which is based on chromium): https://www.srware.net/en/software_srware_iron_chrome_vs_iro...
"Although SRWare claims 'Iron is free and OpenSource' ... Iron 'is entirely closed source and has been since at least version 6'."
Or this:
http://neugierig.org/software/chromium/notes/2009/12/iron.ht...
I also do so with Amazon and Microsoft and would probably claim that I am offering informed consent.
I'm keen to have it explained to me where I might be in error on this matter.
But that's their entire business model.
You might want to pay a lot more attention to SSL-based tracking methods, e.g. https://sites.google.com/a/chromium.org/dev/Home/chromium-se...
Uncoupling your use of Google services from your other online activities is much harder than many people seem to believe.
Visiting a page on a creep's domain to find out about its creepiness seems like a bad idea.
Now that Google has pushed HTTP/2 on everybody, connection reuse will be a major new fingerprinting and tracking technique because it keeps connections open a long time and essentially can't be proxied so one connection is always one browser.
Plus Firefox crashing is a huge downside.
As for privacy implications, there's no reason to believe Google does anything inappropriate, and I rather like having my internet experience tailored based on past behaviour.
To each their own I guess.
How do you get Firefox to crash? What does about:crashes say? Did you report the bug?
Another interesting thing you can do: click through from about:crashes to your crash report. Then click the "more reports" link. You can see if your crash signature is rare or if you have one of the currently "popular" ones. If a lot of other people are crashing with that same signature, there is probably a bug and developers are probably discussing it. Again the most useful thing you can do is figuring out how to make that crash happen consistently and adding steps to reproduce it to its bug.
If that's the case try using 64 bit builds.
Also, pretty much anything that's Javascript + WebGL runs like shit in Firefox. Emscripten demos are good though. But Emscripten runs better in Chrome than JS+WebGL in Firefox.
Compiling all those shaders is rough on any browser, but only Firefox gets its entire UI blocked. Even Internet Explorer manages to isolate the jank to one tab.
Sadly Firefox 39 (nightly) still has no real multi process support. There are only two processes, the main process holds all tabs, the single child process "plugin-container" holds all plugins (Flash, Acrobat, etc.). It's a bit disappointing to see such little progress since 2010. Hopefully Mozilla's Rust based Servo engine comes to the rescue soon and replaces gecko and its old XUL (XML based UI framework) with Servo and an HTML5 based UI (like Firefox OS).
I use Firefox as development browser because of Firebug and and the new developer tools. Working with dozens of tabs open for days is only possible with IE11 and Chrome.
That is incorrect. There is one plugin-container for each plugin module. All Flash instances share a single plugin-container, all Acrobat instances share a single plugin-container, and so forth.
As for the e10s content process, for the moment there is only one content process. The plan is to get e10s working well with the one content process, and then start scaling that up to multiple content processes. Nightly is a bleeding edge development channel, after all.
Instead, I've switched to Disconnect search. They proxy searches and claim to do so anonymously. So at least I'm not directly contributing to Google's profile building that way. It does add a second or two latency per search, which can be annoying.
FF switched to Yahoo, and I'm just amused at how bad it is. They even open results in new tabs by default, a lame trick used to try to accidentally keep people on your site.
It's unfortunate, but Google really does dominate in search.
But mostly I see no benefits from having it enabled, but quite a few negatives (security, privacy, etc). As long as the browser keeps the history of your current browser session alive while the browser stays open, that's perfectly fine for me.
And I agree 100% with you!
I'm a goddamn software developer, software doesn't usually make me feel stupid. So congrats to Google on that I guess...
Storing bookmarks as *.lnk as with IE3+ is still the way to go. Though, I found out that 50% of my bookmarks are obsolete and many were never archived in archive.org due evil robots.txt entries.
Speaking for myself, I always disable (i.e. "rm -rf") auto-updaters because they don't separate security updates from other random changes. Browsers are the worst, with a "major version upgrade" every month dropping whatever the developers had on their build server onto my computer.
http://neugierig.org/software/chromium/notes/2009/12/iron.ht...