1. 29
  1.  

  2. 33

    When content recommendation becomes the most important highlight of a privacy-friendly browser’s new release.

    I love Firefox, but sjeesh

    1. 13

      It is infuriating that the developers of a web browser consider it acceptable to implement any “content recommendation” on their program.

      1. 7

        Why? As stated up-thread, browsing data is never uploaded. Content recommendation happens locally only. What is so wrong with this?

        1. 7

          Because it’s a browser. It should empower me to search the internet for the stuff I want to see, not they think I want to see. I gain no user experience whatsoever. It’s a slippery slope downhill from any recommendation system, no matter how privacy friendly they claim it to be.

          1. 6

            Imagine your new FM radio “recommended” which station to tune to when you turned it on. Would it really be any comfort if the manufacturer assured you that that this recommendation had nothing to do with your own preferences, because they don’t know and definitely don’t care? Bookmarks have been part of browsers since the beginning. This is something else.

            People only put up with this nonsense because it’s free-as-in-beer. That’s why I’d be happy to pay for a fork that treated me like a paying customer rather than a set of eyeballs to sell through some convoluted scheme papered over with a lot of patronizing rhetoric.

            1. 4

              Maybe a car analogy is useful for you. It’s as if your car “recommended” which restaurant to go when you drive it on Saturday afternoon, and actually drove you there without asking, until you overrade it. A minimal amount of fuel would be lost at the beginning of the journey; this is no problem, you can override it at any time. Would you be OK with that?

              I do not want my web browser to make any network request when I open it, unless I ask for it explicitly. As other posts in this thread explain, this is actually impossible with firefox. This is what infuriates me.

              1. 0

                Though they say the browsing data is never uploaded, it’s trivial to match the IP and the time of when the content was recommended. That info can then be correlated with the site serving the recommended content. Several ways exist in which to deanonymize browsing history.

            2. 10

              One of the first things - besides installing uBlock and friends - I do with a new FF installation is the disabling of all these spurious ‘services’ - content suggestion, dangerous content warnings, the various telemetry bits apart from bug reports, those I do send seeing as I run nightly and as such can provide useable reports.

              1. 4

                I would be grateful if you could share the configurations that you are doing. I am asking so that I could note them down and set them too.

                I would like if NixOS would provide firefox configuration options that could be configured centrally, or per user, to make sure that every upgrade applies them.

                1. 3

                  home-manager allows you to declaraticely configure which Firefox add-ons to install (although you still need to enable them manually the first time you start Firefox for security reasons). And you can set Firefox options declaratively using their enterprise policies.

                  1. 2

                    I don’t use much magic to configure it, most is by hand. The only ‘automatic’ thing I do is install a policies.jsonfile (in distribution/policies.json in the FF install directory, in my case that is /opt/APPfirefox/bin/distribution/policies.json) which disables automatic updates since I handle those using a script. I do not want to have binaries writeable by users so these automatic update policies are out of the question. The update script pulls new nightlies from the server and installs them, installs the policies.json file in the correct location and sets ownership and permission so that regular users can execute, but not modify the distribution. I used to have a FF sync server when that was still a thing but eventually it got too hard to reinstate ‘old’ sync support. I do not have, nor do I want to have a ‘Firefox account’ since I do not use any such external services if I can in any way avoid them. I might look into building a ‘new’ FF sync server some time but other matters are more important for now. Until such time I will simply install the following extensions:

                    • uBlock origin (set to ‘expert user’ mode)
                    • uMatrix (disabled by default)
                    • Nuke Anything (to get rid of annoying overlays which uBlock can not filter out)
                    • Open With (to open e.g. media files through a local script)
                    • Containers with Transitions (to always open certain sites in site-specific container tabs)
                    • Foxyproxy Standard (disabled, sometimes used to redirect sites through a local Tor node)
                2. 13

                  It seems to me it’s just a “here are the most popular articles”-list; don’t see anything wrong with that, or any fundamental privacy-concerns. Also from the expanded announcement on it:

                  Recommendations are drawn from aggregate data and neither Mozilla nor Pocket receives Firefox browsing history or data, or is able to view the saved items of an individual Pocket account. A Firefox user’s browsing data never leaves their own computer or device.

                  And from the FAQ:

                  [N]either Mozilla nor Pocket ever receives a copy of your browser history. When personalization does occur, recommendations rely on a process of story sorting and filtering that happens locally in your personal copy of Firefox.

                  1. 11

                    I see something wrong with that, that being giving the user an experience that they have not ask for nor had any control over. Also, what news site, and what collection of news given to the user is trustworthy in a general sense?

                    I feel about it as if I got public broadcasting in my new tab, not something I want nor I am interested in.

                    1. 9

                      that being giving the user an experience that they have not ask for

                      How can you be so sure? I’m a Firefox user, and I find those articles occasionally useful.

                      nor had any control over

                      You can switch it off easily in preferences or directly on the New Tab page (three dots in the upper right corner).

                      1. 4

                        “nor had any control over” is a terrible way to word it (it is your computer and you are definitely in control). My first reaction was “this person is entitled as heck”.

                        However, there is an implied social contract (because firefox existing makes it socially/politically almost impossible to get an alternative off the ground). I still disagree with lich, but their argument has legs.

                      2. 3

                        I see something wrong with that, that being giving the user an experience that they have not ask for nor had any control over. Also, what news site, and what collection of news given to the user is trustworthy in a general sense?

                        I don’t feel that’s a fair characterization. Any new feature can be described as giving the user an experience they did not asked for. And as other commenters note, it can be disabled. Which grants control.

                        As to a user experience, I have lobsters show up in my recommended list, probably because I visit it so often. It does make some sense that I would be recommended what I like to habitually visit.

                        I even removed the suggestion a few times and timed how long and how many visits made it reappear. For me, it learned the association in a day and ten visits to the front page because my habit is to close the tab after quickly reviewing the stories posted.

                      3. 6

                        Where does the aggregate data come from?

                      4. 5

                        I cannot use Firefox and feel safe without ghacks user.js. It is kind of absurd that there is no real community-lead option for browsers. You could put the blame on standards bodies for creating bloated standards, but now more than ever they are just a facade commanded by corporate interests. I don’t know much about it but Project Gemini (along with gopher) seem to be closer to achieving the goals of free software and the “original dream of the web” (whatever that means).

                        Edit: typo

                        1. 1

                          I totally get your point. Would something like Pale Moon feel better to use?

                      5. 10

                        Pocket will now provide fascinating reads from trusted sources in the UK on new tab. I wonder why did they kill RSS support if they intended Firefox to become news aggregator and reader.

                        1. 13

                          Those two are unrelated decisions. The RSS code was bitrotting and that is why it was removed. Pocket is a bundled addon.

                          Yes, they could’ve shipped a bundled addon for RSS as well. But there are a gazillion RSS addons for Firefox that do a much better job than the built-in feature ever did. They are a single click away.

                          1. 4

                            They didn’t need to destroy RSS bookmarks.

                            1. 5

                              They explained the reasons for removing it and there are a dozen add-ons that bring that feature and more back.

                              1. -1

                                Removing support for something doesn’t mean you deliberately destroy user data.

                                1. 7

                                  From the linked article:

                                  When we remove live bookmarks, we will:

                                  • Export the details of your existing live bookmarks to an OPML file on your desktop, which other feed readers (including ones that are webextensions) support importing from.
                                  • Replace the live bookmarks with “normal” bookmarks pointing to the URL associated with the live bookmark.
                                  • Open a page on support.mozilla.org that explains what has happened and offers you options for how you could continue consuming those feeds.

                                  So, the data was not destroyed, it was saved in multiple formats including standard formats made to be imported in other solutions. Documentation was provided to guide the transition.

                                  1. 0

                                    So, the data was not destroyed

                                    My data was.

                                    1. 3

                                      I am sorry, it shouldn’t have been. Have you filled a bug report?

                            2. 1

                              “Bitrotting” is never a justification to remove a feature. It worked perfectly fine until they removed it. Now you have to install RSSPreview, so it’s an extra step to get this basic lightweight functionality.

                              1. 15

                                Have you read the link I pasted earlier? The way the live bookmarks were built were not compatible with sync, used ancient firefox stuff that they are trying to get rid of and had no way to set read/unread status of a post which is a basic necessity of a rss reader. They also have telemetry showing few people using it and more people using more complete add-ons. Mozilla is making a huge effort trying to make Firefox more nimble and maintainable, and the feature was removed. I liked that feature too, but when you read what people maintaining that code and making decisions said about it you might understand why it was done.

                                1. 1

                                  The feature I miss is the RSS previews. The justification given in the blog post holds no weight:

                                  the feed viewer has its own “special” XML parser, distinct from the main Firefox one, and has not had a significant update in styling or functionality in the last seven years.

                                  It is illustrative that Mozilla (or whoever wrote this post) thinks seven years is too long to wait before a styling update. This attitude is responsible for the recent degradation of the about:config page, among other things…

                                  1. 4

                                    You understand that in these seven years Firefox went through a major rework? Working towards removing XUL, XCB, and lots of old code. Adding the Rust bits and new components from Servo. Refactoring the whole UI into an HTML5 based UI. Seven years when we’re talking about what happened in the specific interval mentioned is a lot of changes to the browser code.

                                    Also, as I’ve mentioned multiple times, that feature was:

                                    • Incomplete.
                                    • Not compatible with the current bookmark and sync.
                                    • Better served by multiple add-ons.

                                    People complain about bloat all the time in Firefox. Keep saying that X, Y, Z, features should be add-ons. When they remove a feature that the power users were not using (because they were already using better add-ons) and the casual users were not using (because they didn’t know it existed) taking care to provide a migration path, save the data, and offer alternatives, people cry that they shouldn’t have done it.

                                    It is impossible to please you folks. If they ship some new feature, it is bloat. If they remove something that is hard to maintain and broken, then it is a crime.

                                    Also, the feature is one click away in the add-ons portal. Add-ons exist so that browser vendors don’t need to cover every single feature by themselves.

                                    1. 2

                                      Also, as I’ve mentioned multiple times, that feature was:

                                      • Incomplete.
                                      • Not compatible with the current bookmark and sync.
                                      • Better served by multiple add-ons.

                                      I think you may be referring to Live Bookmarks; I’m referring to RSS previews. I don’t believe these bullets apply to RSS previews.

                                      I also don’t see how Firefox would know how many people used the RSS preview function. If they click on a link to an RSS feed, they would be using it without knowing that they were using it. Then suddenly those links become raw XML and they don’t know that they need the RSSPreview addon to render the page.

                                      It is impossible to please you folks. If they ship some new feature, it is bloat. If they remove something that is hard to maintain and broken, then it is a crime.

                                      Converting an RSS feed to HTML is an extremely simple task that provides functionality that people had come to expect with very little code. New features must meet a higher standard of usefulness and code quality because people aren’t already used to them.

                                  2. [Comment removed by author]

                                    1. 5

                                      If you stop to check Firefox source code and the changes they’ve been doing to “modernize” it away from XUL and many other older technologies inside the code base, you’ll notice that that source code was not up to date with how bookmarks work, it didn’t contain the RSS features people expected it to have (such as read/unread status), and was incompatible with sync and missing from the mobile browsers. That is why it was removed. Read the article, talk to people before spreading FUD.

                                      The feature didn’t disappeared overnight without annoucing and planning a transition. There are a dozen RSS readers on the add-ons portal, all of which are better than that feature ever was. The feature never worked well, that is why it was axed.

                                2. -1

                                  That argument doesn’t support the conclusion. Mozilla removed the RSS code because it was bit rotting: ok, now why was it bit rotting? If the answer was “because it was difficult to monetise” then the parent statement is still true.

                                  1. 6

                                    Since when Mozilla focus feature-wise has been monetization? The RSS code was removed, as explained in the article that apparently no one read, because it was not compatible with the new features they added. The code was bitrotting because it was difficult to maintain and the target userbase moved to add-ons or services with more features. The feature was also incomplete from the point of view of those who make the browser and instead of devoting resources to it, when suitable add-ons are present in the ecosystem, they opted to remove it and focus on other tasks.

                                    Not all features are eternal. Features get removed. This was announced, a path of migration was documented, the data was saved in multiple places. You could easily either move to a SaaS, or to an add-on and import all the same feeds from the OPML export.

                                3. 12

                                  I guess that RSS is more difficult to monetize. They make money on Pocket using ads.

                                  1. 6

                                    Read the article, talk to people involved, go to mozilla community/slack/matrix and talk to people, and you’ll realize the feature was removed for many reasons none of which is monetization.

                                    1. -1

                                      From the article:

                                      Looking forward, Firefox offers other features to help users discover and read content.

                                      Seems to have plenty to do with monetization.

                                      1. 5

                                        The pocket thing is related to monetization. What I’m talking that is not related to monetization is the Live Bookmarks removal that happened in Firefox 64, which was released in 2018. All I’ve been replying and chatting here is related to that feature, and not pocket. The Live Bookmarks removal happened more than 10 versions ago and was not motivated by monetization.

                                        1. 0

                                          The quote above was part of the reasoning behind removing RSS support. In other words they wouldn’t have removed RSS support if there weren’t monetized alternatives like Pocket and sponsored content on the home page.

                                          1. 2

                                            They removed live bookmarks in 2018.

                                            The RSS reader was removed in 2018 alongside the live bookmarks feature. They are part of the same source. Stop spreading FUD.

                                            Pocket sponsored content, which I don’t like either, is not related to the RSS feature.

                                            1. 2

                                              What “other features to help users discover and read content” was the quote referring to?

                                              1. 0

                                                So they destroyed my data in order to monetize me?

                                                1. 2

                                                  No, they told you the feature was going to be removed with months in advance. Provided docs on how to to take your data out, and links to alternative solutions for the feature. Your data should have been saved into two different backups: one OPML on your Desktop, and the bookmarks should still be there, they’d just be normal bookmarks instead of live bookmarks.

                                                  If this didn’t happen, then it is a bug and even though I am really sorry for you, it is the case that the only possible solution is to fill a bug report and try to figure out why the documented process failed so that it doesn’t happen to someone else.

                                  2. 13

                                    I would financially support a fork of Firefox with all the garbage removed. The home page bullshit, Pocket, all the automatic requests made without user action, and more that I probably don’t know of.

                                    1. 8

                                      At least Firefox does allow us to disable anything we don’t want/need.

                                      1. 15

                                        Not everything. Try running Wireshark before you open Firefox and see if you can configure it to produce zero requests on startup. I tried that last week and there was a request to firefox.settings.services.mozilla.com that I think there’s no about:config setting to disable.

                                        Also, there is the fact that when opening a fresh install of Firefox for the first time you’ll make requests to a bunch of random companies, like Facebook, Google etc. The only way to protect yourself from that is to know it beforehand and disable it via user.js before opening the browser. I don’t think that’s right and I wish I could prevent my browser from telling these companies about me. These things should be opt-in in my view.

                                        1. 7

                                          If you disable your internet connection before launching Firefox for the first time, would that give you an opportunity to turn everything off without making extra requests first? I agree that it’s crappy that this is even necessary, but maybe this is a simpler workaround than tweaking user.js?

                                          1. 1

                                            Good idea!

                                        2. 9

                                          The megabar issue proves the contrary.

                                          1. 5

                                            https://www.userchrome.org/megabar-styling-firefox-address-bar.html#mbarstyler

                                            This is not an acceptable solution, by the way: like many organizations, Mozilla needs to get better at incorporating feedback and reverting bad changes. But you did claim it was impossible to undo the megabar, when it’s actually perfectly doable. I also figure that a lot of people would be interested in knowing this.

                                            1. 2

                                              You don’t even need a userChrome for this, there’s an about:config flag (browser.urlbar.update1 I believe).

                                              1. 3

                                                This flag got removed in v77.

                                              2. 1

                                                I did not stay it was impossible, but they aim to make changing it as hard as possible. I do not know the reason for Mozilla acting so about such a minute issue.

                                            2. 2

                                              This isn’t correct. Look at what’s happening with forced updates, for example. Right now I keep installs performed as Administrator (and run as regular user) so the user can’t update, which results in Firefox throwing up dialogs complaining that it can’t update itself. I think I need to go back to building from source just to remove this garbage, but even building from source is more convoluted than it used to be due to sprawling dependencies.

                                              1. 3

                                                You made me realize that this is why Mozilla has zero interest in making an Electron equivalent with Servo.

                                                A Firefox-like browser without the Firefox corporation would be a huge hit and thus unmonetizable.

                                            3. 6

                                              Me too, very much so. Mozilla jumped the shark years back and their attitude does not seem to be improving. Its flagship really deserves a long trip through a detox and weight loss boot camp. It would be a significant effort, though; here are some interesting case studies from Cliqz (RIP) and ungoogled-chromium.

                                              1. 1

                                                I’d really like to see the Tor-browser without Tor. That seems to move in the right direction and they seem to be able to keep Firefox under control somehow.

                                                1. 1

                                                  Why RIP? Cliqz seems to still be active.

                                                  1. 2

                                                    No, it died, the mothership can’t fund it any more. The servers still run, for the time being.

                                                    It’s a pity.

                                                    1. 1

                                                      Holy hell wtf.

                                                      Damn.

                                                2. 3

                                                  There is IceCat [ https://www.gnu.org/software/gnuzilla/ ], though it’s based on ESR releases.

                                                3. 6

                                                  Now I can watch videos in the Flatpak version again! They finally fixed this issue.

                                                  1. 7

                                                    Why do they simply not listen to their userbase? They’ve lost vast majority of the market over time, and just commit mistakes. Last time megabar was extreme joke, and they still have not resolved it.

                                                    I seriously consider moving to a vulnerable mess that is palemoon, or just use chromium-like for webapps and links in graphical mode for most of text based content (which is more than you expected).

                                                    1. 19

                                                      Why do they simply not listen to their userbase?

                                                      Which part of the userbase? The silent majority? Those complaining about trivialities? The few niggled by any change?

                                                      Perfection is ultimately in the eye of the beholder. I’m content to have a real alternative to Chrome, that doesn’t suck, and can be configured to most people’s needs.

                                                    2. 6

                                                      These comments are a sad state of unjustified rage..

                                                      1. 3

                                                        This may be the most stupid thing ever but I just don’t use Firefox more because it doesn’t support macOS dark mode natively. There’s a bug on bugzilla opened 2 years ago and still no action was taken. In addition to that, I think Firefox has some unnecessary features. Even Chrome feels cleaner right after installing. Despite all of that, I like Mozilla’s support of the decentralized web and their take on extensions.

                                                        1. 3

                                                          What’s missing exactly? The default* theme has automatic dark mode and web pages with the correct CSS work perfectly fine as well.

                                                          * sadly, it seems to be the only one. Though it can be modified with userchrome.css

                                                          1. 2

                                                            The native UIs (context menu, save, other dialogs) all use the light theme from macOS. It kinda feels out of place, you know? All other applications I use fit well into the operating system but Firefox (https://imgur.com/a/T5WkXbw). I believe it’s just a matter of using the latest SDK and updating a line on the plist file. At least according to what I’ve read.

                                                            1. 3

                                                              Oh! I’m suprised to say I’ve never noticed that. Now I guess it’ll irk me until it’s fixed.

                                                              1. 2

                                                                Once you see it, you can’t unsee it! 😂