im in ur bookmarks, deleting ur data
September 24, 2018 5:32 PM   Subscribe

Firefox 62.0 removes bookmark descriptions from the UI. Go to Bookmarks, Show All Bookmarks. Notice anything missing? Firefox 62.0 has removed the "Description" free-text field. This has been in the works for some time, as the long-standing existence of this data was flagged as a bug about a year ago. Currently, descriptions are only hidden in the UI; current plan is to delete them from the user's bookmark database altogether in Firefox 64.0 on December 11, 2018.

Hat tip to this AskMe answer. If you use bookmark descriptions, you can use the "Export as HTML" function to save your data and make descriptions visible, though not directly editable. If there's data in there that you need, export it now before Firefox 64 deletes it completely without warning or recourse. This change is necessary because having an extra text field in a database in the year 2018 is an unacceptable burden.
Another option is to use the Firefox Extended Support Release, currently at 60.0, which doesn't include this change yet.
posted by Huffy Puffy (124 comments total) 11 users marked this as a favorite
 
I was ready to be angry and I started to be and then I got to the bottom of the linked bug report.

“I think you misunderstand the nature of the problem, I can assure you we do understand it, after 10 years of maintenance of this specific code. This is not about removing the description field, it's about removing the whole architecture behind it, that was misdesigned and exists just to cover this feature that can be easily (and better) be reimplemented in an add-on, or in Firefox itself, but with a better architecture.”

And having supported some warty software in my time I can really appreciate that perspective. In the end it seems like they are trying to support the core use in a more performant way: “And yes, this architectural choice can cause multi-second (or even multi minutes) hangs in the UI, plus slowing down EVERY bookmark operation, and when you have users with 80 THOUSANDS bookmarks, you can probably understand what it means.”
posted by scamander at 5:51 PM on September 24, 2018 [14 favorites]


Descriptions are a minor feature of bookmarks, they are fetched from the page at the bookmarking time and can be manually edited by the user


Was this a common thing? I gave up on browser bookmarking after Safari updates borked them like they were iTunes libraries.
posted by ChurchHatesTucker at 5:59 PM on September 24, 2018 [1 favorite]


I'm so hardcore I delete the bookmark name too. All I have is a line of favicons in the bookmarks toolbar, and that's how I like it.
posted by scruss at 6:13 PM on September 24, 2018 [32 favorites]


80,000 bookmarks ain't nothin'. I got 80,000 tabs. Take that Firefox!
posted by M-x shell at 6:15 PM on September 24, 2018 [8 favorites]


In my FF it has Name/Location/Tags/Keyword. So what of those is being removed?
posted by cashman at 6:16 PM on September 24, 2018


I get the argument behind it, but it would be nice if they could warn people and/or automatically export the info before deleting the data entirely. (And I don't even use this feature.)
posted by Rhaomi at 6:17 PM on September 24, 2018 [7 favorites]


Or just sign up for Pinboard.
posted by PhineasGage at 6:30 PM on September 24, 2018 [11 favorites]


I use the descriptions for practically everything I take the time to bookmark unless the link name itself is sufficiently descriptive to cause me to remember why I bookmarked something in the first place.
posted by bz at 6:32 PM on September 24, 2018 [5 favorites]


I read that same comment about “misunderstanding the nature of the problem” and it just left me more annoyed. It seems deeply disingenuous to argue that the only two choices are to delete user data or maintain the old architecture forever. That same comment even says they could easily reimplement the feature in Firefox itself. They have have the option of migrating user data to a new system!

I’ve worked on some steaming piles of tech debt, too, and migrating data from the old system to a new one can be a pain. It might not be a good trade-off in terms of time spent, but we don’t know because they ignore that option.

It sucks when you have a bug that loses someone’s data. But deciding to delete their data without even a proactive warning is something else. That’s just so user-hostile, and the way it’s approached here feels like trying to sneak that decision in by the back door.
posted by duien at 6:34 PM on September 24, 2018 [26 favorites]


Sorry, I just read back through my comment and realized it might sound like I’m mad at your comment, scamander. I’m only mad at Mozilla.
posted by duien at 6:38 PM on September 24, 2018 [2 favorites]


I'm so hardcore I delete the bookmark name too. All I have is a line of favicons in the bookmarks toolbar, and that's how I like it.

I thought I was the only one.
posted by pemberkins at 7:15 PM on September 24, 2018 [6 favorites]


That’s just so user-hostile, and the way it’s approached here feels like trying to sneak that decision in by the back door.

Mozilla keeps pulling this shit. If people didn't learn the last time they broke user profiles and lost data, or the last time they sunsetted features used by a whole swath of add-ons, people probably won't learn now. To me, Chrome is like democracy: the worst system in the world except for all the others.
posted by tclark at 7:25 PM on September 24, 2018 [7 favorites]


And Mozilla could be (and has been) brilliant. But they need new leadership post haste.

And I volunteer, no shit. There are not a whole lot of major companies or groups where I personally and honestly believe I could do a better job than the incompetents that run the joint. That list includes Twitter and Mozilla.
posted by tclark at 7:27 PM on September 24, 2018 [1 favorite]


To me, Chrome is like democracy: the worst system in the world except for all the others.

I guess this is timely: "From now on, every time you log into a Google property (for example, Gmail), Chrome will automatically sign the browser into your Google account for you."
posted by MikeKD at 7:32 PM on September 24, 2018 [24 favorites]


Was this a common thing?

There are still a handful of features like that still rattling around in Firefox from the old days - things almost nobody uses, but the people who use them at all care about it a lot.

We do our best to make sure that functionality like this - badly architected, legacy holdovers from the bad old days, but living in the core of the product and something we need to clear out so we can make progress - can be replaced with things like feature-comparable webextensions whenever possible, with hopefully minimal impact.

In this case, this difficult-to-instrument feature was a major stumbling block in our progress towards fully removing blocking main-thread IO, a significant cause of UI hangs in Firefox, and fixing that will be a big improvement for everyone who uses Firefox. I think we definitely could have communicated this better, but I also think this was definitely the right call.

To me, Chrome is like democracy:

Kinda lolled a bit tbh.
posted by mhoye at 7:38 PM on September 24, 2018 [28 favorites]


I got fairly pissed off at a recent "we're killing old code no one wants to maintain" Mozilla move. This one feels less significant to me personally, but I'm sure there's a set of users just as irritated.

After the last thing and a slew of other missteps (some minor, some less so), I think I've finally more or less given up on Firefox. I don't, by that, mean that I've migrated to another browser - the options are for the most part all obviously fucked - or that I wish for the project's failure. I just, you know, pretty much despair. This has been my attitude towards the web generally for a while, so I guess it makes sense that my attitude towards all of the available client software matches.
posted by brennen at 7:44 PM on September 24, 2018 [9 favorites]


I guess this is timely: "From now on, every time you log into a Google property (for example, Gmail), Chrome will automatically sign the browser into your Google account for you."

Yeah Chrome is basically a surveillance tool these days. Which is not to say that other browsers aren't either, but Google can pay their people more, so I assume that the surveillance innfrastructure they get for that is better quality.
posted by carter at 7:50 PM on September 24, 2018 [7 favorites]


idk for all the recent things they've fumbled on, I'm pretty stoked about that "Remove the overlay from this page." button up on the address bar that kills like 97% of modal interruptions. It's the fat X inside the circle. That thing is awesome (I'm on 63.0b8 if it matters).
posted by glonous keming at 7:50 PM on September 24, 2018 [3 favorites]


Q: bookmark folders are still OK, yeah?
posted by eustatic at 7:55 PM on September 24, 2018 [1 favorite]


I got fairly pissed off at a recent "we're killing old code no one wants to maintain" Mozilla move.

Huh, I thought we advertised that change reasonably well, but it was pretty much the same story. Not "nobody wants to maintain this", but "this code can't remain in-place without being a significant drag on progress, and would be better moved to a nice, safe (contained, well-scoped) home in a webextension". The livemarks webextension is the one we ginned up, but I think there's a couple of others.
posted by mhoye at 8:04 PM on September 24, 2018 [4 favorites]


Where does all this anger come from? Big props to mhoye for so gracefully handling dissatisfied comments...
posted by PhineasGage at 8:12 PM on September 24, 2018 [10 favorites]


I'm so hardcore I delete the bookmark name too. All I have is a line of favicons in the bookmarks toolbar, and that's how I like it.

That's... kind of great? I would file a feature request for a option that does that, to make the process less manual and let you keep the favicon but still have mouseovers still show the title. You know about pinned tabs?
posted by mhoye at 8:12 PM on September 24, 2018 [1 favorite]


mhoye, I'd specifically contend (and have elsewhere) that exiling stuff like feed discovery to extensions is itself harmful, but there's no real point re-litigating this kind of thing. I wish y'all well, but I personally give up. The web's a dead loss for structural reasons considerably bigger than my fiddly preferences about long-obscured browser features.
posted by brennen at 8:13 PM on September 24, 2018 [7 favorites]


I think the browser most like democracy is Lynx: ostensibly open to anyone, but difficult to understand and use effectively; subject of bitter debates, but really only very few people keep up with new developments and argue with each other; it was designed back when the world was a dramatically different place, but works with underlying principles you still see everywhere today; and yet in practice it's less effective with every passing year.
posted by shapes that haunt the dusk at 8:33 PM on September 24, 2018 [14 favorites]


Where does all this anger come from? Big props to mhoye for so gracefully handling dissatisfied comments...

I installed LastPass a month or two ago and use it now. Before that, I’d put username and password reminder hints in the bookmarks description field. Now that information is gone with no warning and for no perceptible reason. Now, I happened to find out on Ask MetaFilter that I can export the bookmarks list to html and see it that way, and I think I have most of my important passwords in LastPass now anyway (and lots of them were saved in Firefox, too) but if I didn’t I could have gotten locked out of stuff.

As Jason from The Good Place says, "Any time I had a problem, I threw a Molotov cocktail and, boom, I had a different problem."
posted by Huffy Puffy at 8:34 PM on September 24, 2018 [5 favorites]


To me, Chrome is like democracy:

Kinda lolled a bit tbh.


Yeah, I lolled when Mozilla took over everyone's browser as a stunt to promote the new season of Mr. Robot, too.
posted by tclark at 8:38 PM on September 24, 2018 [4 favorites]


mhoye - I know you mean well, I know this is a project near and dear to your heart, and I'm sure you have an insider's perspective that the rest of us (I assume) don't. But I also sincerely hope you have seen and comprehended the level of anguish that Firefox's users have been experiencing for the past few months. I hope Mozilla is aware that some of the abrupt changes associated with Quantum have been utterly catastrophic to what I suspect is a significant segment of their loyal (up to now, at least) user base, and that they understand how very real and very deep our frustration is. Certainly there's got to be progress, and improvement isn't always a perfectly smooth process. But lately I've been seeing (and experiencing personally) a steady erosion of trust and goodwill and willingness to fight for the underdog on the part of your users, due among other things to how so many of these unwelcome surprises were just kind of sprung on us, and how unwilling Mozilla appears to be to reintroducing (or developing the API features that will allow extension developers to reintroduce) some of the begged-for features that got gutted in the upgrade. You say above that you thought you adequately advertised some of those changes but please be aware that many of us don't feel like we were adequately warned. That deserves measured consideration on Mozilla's part.

I've been holding out on upgrading to Quantum, hoping that the bulwark of defenses I've built up through a combination of hardcore privacy settings, security extensions, and 3rd-party anti-virus/anti-malware utilities will be adequate to compensate for a lack of security/bug updates while I cling to version 56 come Hell or high water - or until Firefox stops actively working against my effort to construct a browser that's safe but that also meets my needs. I've tried upgrading a couple of times on my lower-impact "test" system but the necessary adjustments and the loss of functionality are simply too disastrous, it's absolutely a no-go in my work systems. I can't do it and still continue to do my job. And I hate that an application I've come to rely on to be accommodating to individual users' needs now appears to be diligently throwing that legacy away baby-with-bath-water style.
posted by Greg_Ace at 9:09 PM on September 24, 2018 [14 favorites]


Where does all this anger come from?

We have this conversation a lot, internally; it's part of the price of being an open, transparent organization. I spend a lot of time repeating that most of it doesn't come from people who hate us; it comes from people who love us and are scared that we might be losing our way.
posted by mhoye at 9:10 PM on September 24, 2018 [18 favorites]


As a software engineer, I know all too well how much effort it costs to keep features running. It sucks to have to do it, but any time you have limited resources, you have to make choices about where those resources are spent. I feel for anyone who used this feature, but sometimes that's how it has go. I've had to re-write a Firefox extension when it moved to Quantum and the new Web Extension API that went with it, so I know how painful it is to lose features you're used to, but the night-and-day difference in speed I got from Quantum was well worth it, and even though it hurt, I know they made the right choice pushing that forward.

I suspect that my earlier-drafted comments about people kvetching about a browser that they downloaded for free would probably not go over well, so I won't go into that more than to encourage people to donate to the Mozilla Foundation if they're so inclined. (Note: I'm not affiliated in any way with Mozilla, I just like Firefox.)
posted by Aleyn at 9:13 PM on September 24, 2018


Where does all this anger come from?

You think people are mad now, wait till they come in to work one fine December morning and discover Mozilla has deleted a bunch of their data.
Not made inaccessible within the UI but removed from the product entirely.

Whoever made that decision needs to have their head examined.
posted by madajb at 9:24 PM on September 24, 2018 [9 favorites]


Maybe it’s just me, but I’ve been using Firefox since it was the Phoenix beta and honestly I can’t see a single thing about Quantum that I can complain about. It’s fast, it does what I need, and I’m happy. I’m not sure what others are missing with recent updates. I certainly don’t miss how slow it used to be. Removing features that are underutilized and cause major problems - especially when there are workarounds - that doesn’t seem user hostile. It does seem weird to expect the demands of a few vocal people to overrule the experience of the majority of users. As much as it sucks to be on the negative end of a UI decision, I can’t get worked up about this one.
posted by caution live frogs at 9:24 PM on September 24, 2018 [4 favorites]


I think people would be a lot more receptive to the "just do it in an extension" argument if browsers didn't have a history of progressively limiting the capabilities of extensions.
posted by Pyry at 9:28 PM on September 24, 2018 [18 favorites]


I think the browser most like democracy is Lynx: ... and yet in practice it's less effective with every passing year.

Note that we're not talking about Lynx' bookmarks.
posted by ChurchHatesTucker at 9:34 PM on September 24, 2018


My desktops always have Firefox with noscript.

I use safari and chrome (only for my old-guy yahoo mail account) and...

My bookmarks are mistakes most of the time. Recent history or I've likely moved on. I'm not going to "check that later". POTUS thread and all...
posted by Windopaene at 9:35 PM on September 24, 2018


> "But I also sincerely hope you have seen and comprehended the level of anguish that Firefox's users have been experiencing for the past few months."

Not months. It's years now; not too far off a decade since v4 & Aurora were the first signs of Mozilla mistaking 'contributors' and 'bugzilla commenters' for 'users'.

I've been using Firefox since it was Phoenix v0.2. I've grudgingly moved as far as 52.something.something ESR, despite multiple near-dealbreaking functionality changes, but have avoided going further because of actual genuine dealbreaking functionality changes.

Despite its own many many problems, I've about 50% moved to Safari. It doesn't have a lot of the features I use daily. Yeah, it doesn't have any sort of user-annotation for bookmarks. Yeah, in the future it probably will remove a few features it has that I do use.

But I don't expect better of it. With Mozilla, I did…
posted by Pinback at 9:46 PM on September 24, 2018 [2 favorites]


Reading the bug, it sounds like the big sticking point here is that the description field was auto-populated with whatever a page's HTML description contained, so now there's no way to automatically distinguish between the 99% of users who have never edited a bookmark description in their lives, and the people like Huffy Puffy who, unbeknownst to everyone, happen to have organized their lives around the feature. If there were some way to identify people who have actually used the description field, it would be possible to let those users know it's disappearing and export their data for them (or prompt them to use a suitable extension).

But since the feature was badly designed from the start, there's no good way to save just the useful data for the people that used the feature without giving everybody else a big pile of junk.

But the outright hostility to even acknowledging that deleting even a small number of users' data without much in the way of notice is not good. Other solutions aren't great, such as tossing it into a text file in the user's profile, which has privacy implications, but this isn't the right way to handle deleting people's stuff. Fortunately, no data has actually been deleted yet and won't be until a subsequent version.
posted by zachlipton at 9:56 PM on September 24, 2018 [3 favorites]


I think people are used to being pushed around by Google and Microsoft and Apple and they look to Mozilla as a respite from that, whether based in fact or not. So when Mozilla acts in a way that's perceived as authoritarian people get a little more mad or sad than they would otherwise because they hope for better. They need to give people the option to save if they don't happen to be Hacker News readers.
posted by M-x shell at 10:02 PM on September 24, 2018 [10 favorites]


It seems like the obvious solution would be to automatically export the descriptions to an html or JSON file as part of the Firefox 64 installation process. That way the data is not lost for anyone who doesn't read the patch notes (because most users don't, until they have a problem!).
posted by JDHarper at 10:57 PM on September 24, 2018 [5 favorites]


If I trust the Mozilla sync servers I expect them to respect my data. Randomly deleting data breaks that trust. I would expect Mozilla to give large and explicit warnings, and a link to a suggested replacement extension.
posted by jaduncan at 10:59 PM on September 24, 2018 [2 favorites]


At least it's not Safari, the most garbage browser to ever garbage.

Says the guy who used to create small AVI clips because Microsoft insisted animated gifs were not going to be used with their new browser, IE 1.0, no sir.
posted by maxwelton at 11:30 PM on September 24, 2018 [2 favorites]


Ok, I've been drinking a little bit so I'm overcoming my normal recalcitrance to comment on computer things on mefi.

Please update your browsers. I would go so far as to say that if you could only do one thing the ensure that your computer doesn't get screwed, it would be update your browser. Updating your browser is more valuable than all your anti-virus or any of your browser extensions to keep you safe. I'm not sure if people realize how frequently all browsers get popped. Someone up thread say they were sticking to Firefox 56. Not to pick on them, but Firefox 56 was released roughly a year ago. Let's use that as a case study

Since then:
Firefox 57: 2 critical bugs
Firefox 57.0.1: 1 critical bug
Firefox 57.0.2: 2 critical bugs
Firefox 57.0.4: spectre mitigations
Firefox 58: 2 critical bugs
Firefox 58.0.1: 1 critical bug
Firefox 59: 4 high sev bugs
Firefox 59.0.1: 3 critical bugs
Firefox 59.0.2: 2 high sev bugs
Firefox 60: 6 high sev bugs
Firefox 60.0.1: 1 critical, 1 high sev
Firefox 61: 3 critical, 5 high sev
Firefox 62.0.0: 2 critical

All from https://www.mozilla.org/en-US/security/known-vulnerabilities/firefox/

This totals to 36 bugs that potentially lead to someone else running code on your computer (the rubric they give for critical vs high sev looks like it changes around Firefox 58, they start listing memory corruption bugs as high sev instead of critical). Any single one of these could lead to crypto miners being installed, ransomware, botnets, or worse. Additionally, these are only the bugs Mozilla has been made aware of! There are almost certainly more.

If this point isn't driven home enough yet, you can literally just search google and find an exploit (safe to click, PoC code) for that version of Firefox. This is obviously not a complete PoC for code execution (I think it only crashes under special conditions), but it's a strong sign in my opinion that you might want to reconsider your approach here.

Look, I also feel somewhat held hostage by the browser manufacturers (and similarly OS manufacturers) and it sucks because your options are a) your computer and data are now everyones computer/data! b) use a product which pisses you off in some way c) not use a browser. None of these are good options, but at least currently thems the breaks. Additionally, I'm aware this is not a popular viewpoint on MeFi, where people will regularly proudly proclaim their of some old software ("macOS 10.6, the last good one they made!"), but the same kind of number apply there as well. Good luck I guess.

Finally, just to be clear this is not picking on Firefox. While I have opinions on the relative security merits of one browser vs another, all of them get exploited with alarming frequency. The point I'm making is that your goal should be to increase the cost of someone else running code on your computer to such a point where it not feasible for average internet schmucks. The easiest way to do this is to update your browser. Please.
posted by yeahwhatever at 11:44 PM on September 24, 2018 [18 favorites]


And yes, this architectural choice can cause multi-second (or even multi minutes) hangs in the UI, plus slowing down EVERY bookmark operation, and when you have users with 80 THOUSANDS bookmarks, you can probably understand what it means.

Oh no, 80,000 strings! That's probably 2-5 whole megabytes!

Just make your architecture less stupid. Seriously, what are you doing, downloading the text from the cloud over a 56k modem? No, you're reading some local text file, probably off a SSD. Read it and cache it in ram. The guy with 80,000 bookmarks probably thinks they're important enough to waste a few megabytes (god knows the rest of your browser eats mem like candy).
posted by ryanrs at 12:07 AM on September 25, 2018 [3 favorites]


mhoye is a saint.
posted by nnethercote at 12:36 AM on September 25, 2018 [8 favorites]


Please update your browsers.

Please provide browser updates that continue to do the things I need to do.
posted by straight at 1:10 AM on September 25, 2018 [7 favorites]


I'd like to hear more about why storing and serving the bookmark descriptions field was such a pain. Not because I think the good people giving me free software need to justify themselves, just because it sounds interesting. Was it something to do with thread blocking?
posted by Joe in Australia at 3:28 AM on September 25, 2018 [1 favorite]


> In my FF it has Name/Location/Tags/Keyword. So what of those is being removed?

None of those, it's a (now invisible) text field users could write into and that sometimes was filled with a mostly useless auto-generated blurb from the bookmarked site.
One would have to go out of their way to use it, afaik it required bookmarking a site then explicitly opening the bookmarks library and editing the description.
posted by Bangaioh at 4:47 AM on September 25, 2018 [2 favorites]


Anything they can do to stop Firefox from hanging at key moments is okay by me. I have too many handy plug-ins that I like and use to ditch it for Chrome, but it's showing its age.

I sometimes export my Bookmarks to HTML so that I can grab links with their code to use in my blog, and just to back them up locally on occasion. Is this field the one that's stored between <DD> tags? That one's almost always full of auto-generated text, which could add up over thousands of bookmarks. (The other dead weight in an exported bookmarks file is the code for every single favicon—the part that reads ICON="data: ..."—stripping those out with a regexp search-and-replace cuts the size down enormously.)
posted by rory at 5:23 AM on September 25, 2018 [1 favorite]


Please provide browser updates that continue to do the things I need to do.

For free!
posted by srboisvert at 5:24 AM on September 25, 2018 [3 favorites]


@srboisvert - "For free!"

This is really only a fair dig when, you know, users have an option to pay for something.

The options, as I see them, for normal humans who don't want to write their own browser from scratch are:

- OS-bundled browsers with macOS (Safari) or Windows (Edge) that appeal to some folks, but also are strongly tied to those platforms and may not work with all sites since they're not dominant.
- Chrome - which has its own history of disappearing features and inserting user-hostile features. (See current debate about "signed into Gmail? OK, you're signed into the browser too!")
- Chromium that is open source but also suffers from the above sign-in problem as I understand it.
- An assortment of proprietary or quasi-proprietary browsers like Brave or Opera with really tiny user bases.
- An assortment of open source browsers that (in my experience) mean hard tradeoffs in terms of what you can actually do online (e.g., Midori, Konqueror, etc.) and are platform limited.
- Firefox.

The common thread? None of these AFAIK operate on the basis of paying for a browser where users might reasonably have input into the feature set.
posted by jzb at 5:35 AM on September 25, 2018 [3 favorites]


I think people would be a lot more receptive to the "just do it in an extension" argument if browsers didn't have a history of progressively limiting the capabilities of extensions.

I think that if Mozilla was actually serious about "just do it in an extension" and believed they were in fact leaving enough capabilities to do so, they would build in time to this very important architectural refactor that's requiring this feature be gutted for writing an initial version of such an extension and publishing it to a public repo so that people who cared could maintain it. That they don't do this seems like ample evidence that "you can just write an extension" is a convenient way to tell people to fuck off while pretending they're not telling people to fuck off.
posted by tocts at 5:43 AM on September 25, 2018 [5 favorites]


Based on this discussion, I just installed the latest v. of Firefox to my home desktop running macOS 12.6. I'm immediately wowed by how fast and responsive it is compared to Chrome! I haven't used FF in years since switching to Chrome and it feels great.

I think I'll make the switch on desktop. I've tried Firefox mobile and really don't like the 'tab drawer' compared to Chrome's. Chrome's verticle card stack makes it much easier to swipe away tabs one-handed. The way Firefox mobile lays them out in a gride means straining my thumb or using two hands to reach the 'x' on a tab card.
posted by Evstar at 6:01 AM on September 25, 2018 [1 favorite]


I think people would be a lot more receptive to the "just do it in an extension" argument if browsers didn't have a history of progressively limiting the capabilities of extensions.

But, by extension, making them a hell of a lot safer. You limit the capabilities of the extensions to do real harm.
posted by urbanwhaleshark at 6:05 AM on September 25, 2018 [5 favorites]


FWIW, there's really only one last Chrome feature that keeps me there, and that's the multi-user personalization in the same browser process. If FireFox ever gets that without my having to make custom app launcher shortcuts (macOS does not make this easy) I'll probably switch.
posted by ChrisR at 6:05 AM on September 25, 2018 [2 favorites]


I've tried Firefox mobile and really don't like the 'tab drawer' compared to Chrome's.

I haven't looked at it in a while, but I think there's a menu setting for that to get it back to one column. "Compact tabs", possibly.
posted by Mr. Bad Example at 6:27 AM on September 25, 2018 [1 favorite]


Oh no, 80,000 strings! That's probably 2-5 whole megabytes! Just make your architecture less stupid. Seriously, what are you doing, downloading the text from the cloud over a 56k modem?

If you block main thread I/O for more than about a 20th of a second, it becomes a user-perceptible performance issue. "Jank", we call it; as in, the browser gets janky. In practical terms, despite how blazingly fast modern CPUs are, it doesn't take more than a few round trips to a spinny disk or one accidentally-On^2 algorithm baked into the wrong corner of the code for that to show up to the user as something that feels like a crappy product.

And unfortunately, that's where we are here, and "just make your architecture less stupid", when we're working on parts of the codebase that are fifteen or twenty years old isn't really a "why don't you just" kind of question. In a lot of respects this is like mucking around in the foundations of an old building that happens to have a few hundred million people living and working in it every day. But yes, we're making our architecture less stupid. This is an important step in that direction.
posted by mhoye at 6:31 AM on September 25, 2018 [22 favorites]


mhoye, you are (as someone said upthread) a saint. Your patience amazes me.
posted by ChrisR at 6:32 AM on September 25, 2018


> FWIW, there's really only one last Chrome feature that keeps me there, and that's the multi-user personalization in the same browser process. If FireFox ever gets that without my having to make custom app launcher shortcuts (macOS does not make this easy) I'll probably switch.

Does MultiFirefox not fit the bill? It requires relaunching the browser but does not require writing launcher shortcuts.
posted by ardgedee at 6:45 AM on September 25, 2018


It does not. I keep one Chrome profile for my personal use, signed in to my external accounts, and then another for work use that connects to internal resources. Beyond that I usually have one for each project I work on as a hobby where they are associated with different account cookies etc. Right now, I have four profiles open at the same time, and don't foresee that ever dropping below two.

ETA: I'm gonna have a second look at it, though. It might be enough to make this doable.
posted by ChrisR at 6:49 AM on September 25, 2018


I think people would be a lot more receptive to the "just do it in an extension" argument if browsers didn't have a history of progressively limiting the capabilities of extensions.

I talked a little bit more about this here, and I think that argument has held up well; we don't (and won't, and can't) let webextensions do literally anything for the same reason we don't let javascript do literally anything. We've tried that, and it turns out to go pretty badly for a pretty large number of people.
posted by mhoye at 7:12 AM on September 25, 2018 [1 favorite]


(Do you need entirely different profiles, or just different login cookies and the like? If the latter, check out the official Multi-Account Containers addon. It's slick.)
posted by introp at 7:15 AM on September 25, 2018 [2 favorites]


"In a lot of respects this is like mucking around in the foundations of an old building that happens to have a few hundred million people living and working in it every day. But yes, we're making our architecture less stupid. This is an important step in that direction."

I think this is a fair thing to do, feature-wise. Maintaining software over years and decades is hard and you cannot please everyone. You certainly can't tend to every feature that someone loves while maintaining performance and moving forward. I get that.

The thing that I'm hung up on is the deletion of the data, apparently without notice to the user. Is there no way to give the user notice and export the data in case they are one of the (I'm sure few, but non-zero) users who may have something stored in this field that they cannot replace?
posted by jzb at 7:15 AM on September 25, 2018 [4 favorites]


I keep both a Chrome and a Firefox session open. I need Firefox because it allows me to have that sidebar bookmarks menu on the left. Chrome has no idea what that is anymore.
posted by Ber at 7:53 AM on September 25, 2018


Please provide browser updates that continue to do the things I need to do.

@srboisvert - "For free!"


Don't you dare say that when you are the one asking me to throw away a tool that works and put in a bunch of free labor figuring out how to make the new tool do what the old one did.
posted by straight at 8:11 AM on September 25, 2018 [5 favorites]


Greg_Ace: what exactly is keeping you on unsupported releases? If an old extension is that important to your daily work, more important than the security, performance, and reliability benefits of upgrading, are you at least supporting its ongoing development?
posted by adamsc at 8:11 AM on September 25, 2018 [1 favorite]


The thing that I'm hung up on is the deletion of the data, apparently without notice to the user. Is there no way to give the user notice and export the data in case they are one of the (I'm sure few, but non-zero) users who may have something stored in this field that they cannot replace?

I had the same thought. Mhoye, I know it's not your job to justify the entire company's decisions to us at MeFi, but do you have an insight into why this isn't happening? Maybe a modal notice before update that "you need to save your Descriptions RIGHT NOW or they'll be deleted," or an auto-export to JSON if the user's Descriptions are populated...?

Beyond that I also find the entitled anger off-putting. Like, the direction Chrome is going is much more annoying and concerning, but everyone's just like "forget it Jake, it's Chrome-town." I also have a friend in Mozilla legal; I should take him out for a beer and tell him to keep up the good work.
posted by Joey Buttafoucault at 8:13 AM on September 25, 2018 [2 favorites]


So, they've hidden it and they'll delete the feature in January (at the earliest, since that choice could slip to a later release).

Do you all realize that for anyone who uses that feature often this amounts to a three month notice? If I noticed this missing key, today or soon, what would I do? I'd do a google search for what the hell happened and I'd very likely find the release notes that clearly state "Users who have stored descriptions using the field may wish to export these descriptions as html or json files, as they will be removed in a future release." These notes include links to show me exactly how to retrieve my data.

So, this is not some random, malicious, incompetent deletion of data. It's a preparatory step that effectively gets users attention *without* deleting data. The uses then have months to figure out what they need to do to persevere through the more permanent transition.

Perhaps Mozilla could have handled the announcement a bit better (a launched tab when the updated browser first launches?) but they have not hidden the information. It's all right there for anyone that goes looking for it.
posted by oddman at 8:29 AM on September 25, 2018 [3 favorites]


Chiming in to thank mhoye for their explanations and patience.
posted by terrapin at 8:38 AM on September 25, 2018 [1 favorite]


And unfortunately, that's where we are here, and "just make your architecture less stupid", when we're working on parts of the codebase that are fifteen or twenty years old isn't really a "why don't you just" kind of question. In a lot of respects this is like mucking around in the foundations of an old building that happens to have a few hundred million people living and working in it every day. But yes, we're making our architecture less stupid. This is an important step in that direction.

Why didn't you provide a defined migration path to a better-architected version of the feature (or even an extension) before you decided to delete user data? Seems like that would have avoided all the controversy.
posted by cosmic.osmo at 8:44 AM on September 25, 2018 [1 favorite]


nthing mhoye for sainthood. I'm really happy with how firefox is evolving. It feels like a faster and more secure browser with every release, and one which is making a good attempt at protecting me from the growing panopticon.

And yah... the entitlement. I think the world would be a much better place if people stopped assuming that other people's jobs were easy. Or similarly, if people would assume that changes in a product which negatively affect them are the result of making difficult comprises, as opposed to maliciousness/indifference/incompetence. I mean, if you can't believe that a non-profit like mozilla is generally operating in good faith, how do you trust any organization?
posted by Alex404 at 8:46 AM on September 25, 2018 [4 favorites]


Do you all realize that for anyone who uses that feature often this amounts to a three month notice? If I noticed this missing key, today or soon, what would I do? I'd do a google search for what the hell happened and I'd very likely find the release notes that clearly state "Users who have stored descriptions using the field may wish to export these descriptions as html or json files, as they will be removed in a future release." These notes include links to show me exactly how to retrieve my data.

You're assuming that users who have data stored here use the feature often, which may not be the case.

Look, if I use a program and store some data in it, I don't expect that program to delete that data ever without explicit notification. That is a violation of user expectations of the highest order. We're not just talking about a feature going missing, that's their data.

You're also operating with the assumption that a user who depends on this data is good at google search, understands release notes, etc. which may very well not be the user profile of someone who's stashed data in this field.

I don't think that "explicitly tell me before my data is deleted" is an unreasonable position to take.
posted by jzb at 8:55 AM on September 25, 2018 [8 favorites]


I had the same thought. Mhoye, I know it's not your job to justify the entire company's decisions to us at MeFi, but do you have an insight into why this isn't happening?
[...]
So, they've hidden it and they'll delete the feature in January (at the earliest, since that choice could slip to a later release). Do you all realize that for anyone who uses that feature often this amounts to a three month notice?

We do, yeah. We could have done this better, and to the point about "just making our architecture less stupid", this was a place where we've got very little insight about real-world usage of the product. That's part of what we're working to fix here, as well as some tooling to provide people who rely on this field with a migration path that at the very least promises that nobody suffers the indignity of dataloss. "The data is still there" is an admittedly dissatisfying answer, and we're also working through the details of process change to make sure in some depth that even perceived changes to user data are handled with a great deal of care and respect even at the thinnest margins of usage.
posted by mhoye at 9:03 AM on September 25, 2018 [3 favorites]


The thing is, causing user data to be corrupted or deleted is usually considered the cardinal sin for I/O. For obvious reasons.

This specific issue isn't about entitled users whining - this is an appropriate WTF reaction separate from the decision not to continue having this data. I definitely understand (though find unfortunate) the need to remove the functionality now instead of spending the resources to provide it in an alternative way *before* removing it. But I find it really hard to believe that doing an automatic export of users' data before deleting it would take up an unacceptable amount of resources on Mozilla's part.

I love Firefox and care about it partly because the other available browsers are so problematic. But a decision-making process that says "some users will lose their data, but we can live with that" is really not all right.

On preview: glad you're working to prevent dataloss. I really hope this wasn't an afterthought.
posted by trig at 9:06 AM on September 25, 2018 [3 favorites]


And yah... the entitlement.

For what it's worth, I don't think that's fair. This is user data we're talking about, and the fact that users are in fact entitled to control over their own data is a real thing we believe in, and a standard we've set for ourselves and aspire to live up to. We haven't gotten that right here, mercifully not unrecoverably so, but we should be held to account when we don't meet the standards we've set for ourselves, and to learn from that and do better when we don't.
posted by mhoye at 9:08 AM on September 25, 2018 [13 favorites]


"You're assuming that users who have data stored here use the feature often, which may not be the case"

I do assume this. But what time window is reasonable? If a person doesn't use the feature in six months is Mozilla stuck waiting for them? How about a user that hasn't launched Firefox in years but still technically has some annotations in Firefox? Does Mozilla just say to the rest of their users "Sorry we can't improve performance because some people used this feature once and now we have to support it forever."

"I don't think that "explicitly tell me before my data is deleted" is an unreasonable position to take."

So, how should Mozilla contact these people who use the feature, but not very often? E-mail? There will always be some people who miss the notice. That's just how this goes. Getting every single person who (irregularly!) uses a feature to notice that you are changing it is a practical impossible.

While I agree that features ought not to be changed capriciously. This is not simply a decision without an explanation. Mozilla believes that the improvements this will allow are valuable to the whole userbase. At some point sure the benefit to the many outweighs the preferences of the (again! irregular!) few still using an outdated feature.
posted by oddman at 9:09 AM on September 25, 2018 [1 favorite]


I do assume this. But what time window is reasonable? If a person doesn't use the feature in six months is Mozilla stuck waiting for them? How about a user that hasn't launched Firefox in years but still technically has some annotations in Firefox? Does Mozilla just say to the rest of their users "Sorry we can't improve performance because some people used this feature once and now we have to support it forever."

Forever. There is no reasonable time window to delete user data without warning. None. To disable a feature? That varies based on the cadence of the project and user profiles - so for Firefox, deprecating a feature and then disabling it w/in 3 months isn't the worst thing I've ever heard. But deleting data?

Imagine having some piece of data stored in a program that you have on a VM or a computer you rarely use - you power it up, and $program says "I have an update" so you update it. Then you go to look something up and *poof* it's gone. No warning, no opportunity to back it up.

I get that this isn't going to affect tons and tons of users, but I think "never delete my data without warning me" is a completely reasonable position to take no matter how few users are affected.

So, how should Mozilla contact these people who use the feature, but not very often?

I'd expect a dialog or something in the update mechanism that warns people before they take the step of updating.
posted by jzb at 9:15 AM on September 25, 2018 [6 favorites]


For what it's worth, I don't think that's fair.

Yet again demonstrating your excellence at community management ;)

Unsurprisingly this issue doesn't affect me, which is why I'm not incensed by it. I think many of the comments here represent a good faith attempt at pointing this issue out, and making sure that mozilla addresses it in a satisfactory manner. It's only the occasional hostility that I find frustrating (again, not that I have a horse in this race) because I worry that it might demotivate the many people at mozilla who are working hard for their users.

It saddens me that, at least superficially, metafilter seems to produce a lot of negative 'coverage' of firefox. I'm not an idealogue, but I continue to believe that firefox is an important product, which a lot more people should be using. I wish we could generate some more positive coverage (and analysis, constructive criticism, etc...) of the positive direction of firefox, as other browsers are taking the opposite path.

(or maybe I should just make my own post about it, eh?)

Anyway, keep up the good work, is all I'm trying to say.
posted by Alex404 at 9:21 AM on September 25, 2018 [8 favorites]


Yes, please, for the love of god, stop running old versions of web browsers. No combination of security settings is going to help you.

FWIW, I recently switched back to Firefox. From my perspective, the changes introduced in "Quantum" were a huge and substantial improvement in the browser's speed and stability. Firefox is the only browser that will load one of MeFi's gargantuan #uspolitics threads on my phone or old-ass computer without grinding to a halt.

The new Extensions API is nice too. It's now considerably easier to develop extensions that target both Chrome and Firefox, which has massively expanded the selection of extensions available.

It's a shame that Firefox had to break compatibility with old extensions, but it was a very long time coming, and there were a multitude of extremely good reasons for Firefox to rip that band-aid off.

Also, Firefox is local software. Your bookmarks are not in the cloud. Make backups. [I don't really agree with Firefox's decision to eventually wipe this data, but one of the advantages of local software is that, as long as you're making backups, you still have complete control over your data.]
posted by schmod at 9:39 AM on September 25, 2018 [1 favorite]


I would be happy if I could just get Firefox to stop scrambling the favicons on my bookmarks randomly. I've deleted favicons.sqlite and repopulated with fresh favicons, and a couple weeks later, the favicons start scrambling again, displaying the favicon from one bookmark on two or three other random bookmarks.
posted by JackFlash at 9:40 AM on September 25, 2018 [1 favorite]


Yeah Chrome is basically a surveillance tool these days
[citation-needed] -- There's a lot of really bad shit happening on the web these days. Google is responsible for some of it. As far as I am aware, Chrome's behavior has always been fairly reasonable -- the recent update is far more innocuous than it's been made out to be (and, in fact, helps to reduce a potential source of user confusion that could cause you to inadvertently leak data to another user).

Baseless accusations like this one are not helpful. There's a lot of actual-surveillance going on.
posted by schmod at 9:44 AM on September 25, 2018


I apologize for the hostile tone of my comments. I do think it's unrealistic to expect that everyone is going to be willing to trade a tool that works for one that doesn't--but might possibly do the same things if you put a lot of work into figuring it how but nobody will guarantee that--just because the current tool that works might cause them problems in the future.

For the record I finally upgraded to the current Firefox, going along with trading extensions that might be insecure for trying to figure out which 3rd party programs do the same things and are are not malware and which sites will distribute them without bundling malware into the installers. I suppose this is still on balance more secure.
posted by straight at 10:02 AM on September 25, 2018


You know about pinned tabs?

Neat, but not what I'd use. I don't like to keep more than 5-6 tabs open, and some of mine are huge memory hogs.

I'll maybe get round to that feature request. Glad that at least one other person doesn't think it's batshit.
posted by scruss at 10:05 AM on September 25, 2018 [1 favorite]


Welp. Glad I read about it here, because I use that field to enter username and password hints for myself (NOT THE PASSWORD, DUH) for every damned site that demands its own unique credentials.

And speaking of updates, I'd be a lot more compliant with it if they weren't released, like, daily. I'm just not going to deal with closing all my tabs, trying to fix what the update invariably breaks, etc., every damned day; I just don't have that much time for you. I've gotten very used to clicking "ignore" on the update prompt (yes, I see you cringing).
posted by Dashy at 10:05 AM on September 25, 2018 [4 favorites]


So, regarding the "this is user data that might be lost" argument... I could be misreading it, but I think one of the core issues with this field in the bookmarks DB is that in the vast majority of cases it is filled in without user intervention and never interacted with. Informing all users of the deprecation of something that very few users use is a bad UX for the majority of them. The ones who use it will in theory notice its absence in the window where the data is retained but the frontend for it is absent, and since the data is still there they can recover it, even if it's sort of on them to figure out what to do with it after that.

I think it's safe to say that messaging hundreds of millions of users for something that a tiny fraction of them use, with wording that leaves them fearful of data loss for something they don't understand and don't use, will erode more trust in Firefox than the alternatives. I'm not sure how to navigate the two without some cost, and looking to the longer term feels like the right choice here.
posted by ChrisR at 10:13 AM on September 25, 2018 [2 favorites]


From now on, every time you log into a Google property (for example, Gmail), Chrome will automatically sign the browser into your Google account for you

right, but then you just? sign out of gmail? and you are no longer signed into chrome? i guess i just don't understand 01) staying signed into gmail all the time or 02) signing into chrome for anything other than one time bookmark syncs when you get a new computer? what are people doing this for.
posted by poffin boffin at 10:21 AM on September 25, 2018


For some of the applications I use that have an updater, the prior configuration is archived before any changes are made. I think that as this feature is phased out, automatically exporting that data into html or json files should be a priority. Also, I think that sponsoring an extension as an upgrade path is a good idea. And since Mozilla has a channel on the new tab page for pushing news to users, I think changes like this need to be in the rotation above blog posts about data-collection ethics.
posted by GenderNullPointerException at 10:58 AM on September 25, 2018 [1 favorite]


signing into chrome for anything other than one time bookmark syncs when you get a new computer? what are people doing this for.
If you have multiple computers, it's pretty helpful to stay signed-in so sync happens continuously.

Also, a lot of people choose to sync their history. There are some pretty big privacy implications to this if users are commonly confused about which user their chrome is logged-in as.
posted by schmod at 11:46 AM on September 25, 2018 [1 favorite]


poffin boffin: The carrot dangled in front of sync is that everything you do on desktop Chrome will be available in Android Chrome and ChromeOS. In fact, a big selling point behind ChromeOS for education is that you can reassign tablets and laptops to students, and their personalization lives in the Google cloud.

That is, unless you explicitly turn those features off, which isn't as easy or transparent as claimed. Since Google is a known bad-faith actor with "we're not really tracking your location but we're tracking your location telemetry," there's no way to verify that's the case. So the best solution is to just not sign in on Chrome or Chromium.

IMNSHO it violates the principle of least privilege for an HTML5 interface like Gmail, Maps, Drive, or Hangouts to assume implicit permission to access to APIs for browser settings. No, I don't want to give you my metafilter activity (among other things) because I need to find a pizza place.
posted by GenderNullPointerException at 12:14 PM on September 25, 2018 [3 favorites]


I've been with this codebase since it was Netscape Navigator. I learned alot about sed, awk and grep via trying to turn the bookmarks.html file into a blogging engine simply by annotating bookmarks, using this very feature (description), and throwing the result up on another html page on my college account - it didn't work very well. It's sad to see it go, but lots of features have come and gone, and I learn to get over it. I learned what little I know about sql by trying to pull data from places.sqlite3. The only one that really stung was an addon that let you host and sync bookmarks via ssh (I forget the name now), and I really liked it, and was sad to see it go. I guess I'm team 'meh'.
posted by eclectist at 12:31 PM on September 25, 2018


> Is this field the one that's stored between DD tags? That one's almost always full of auto-generated text, which could add up over thousands of bookmarks.

Yes, that's the one.

(The other dead weight in an exported bookmarks file is the code for every single favicon—the part that reads ICON="data: ..."—stripping those out with a regexp search-and-replace cuts the size down enormously.)

1. close firefox
2. delete favicons.sqlite from the profile dir (or temporarily move it out of the way and restore it later)
3. open firefox without net connection (to prevent refetching icons if you have any tabs open by default)
4. export bookmarks without bloat (should be the default tbh)


> Do you all realize that for anyone who uses that feature often this amounts to a three month notice? [...] The uses then have months to figure out what they need to do to persevere through the more permanent transition.

Not enough for me, I hoard a bunch of unsorted bookmarks and a couple of times a year I organise them and add descriptions where needed. I don't read release notes, it's not unconceivable this would have passed unnoticed and that I'd have rotated away my last good backups.
posted by Bangaioh at 1:02 PM on September 25, 2018 [3 favorites]


If you block main thread I/O for more than about a 20th of a second, it becomes a user-perceptible performance issue.

I still don't get why the description is so much more onerous to store and access than the URL itself or the favicon, which you're presumably keeping.

And from a high-level architecture standpoint, once you've accepted bookmarks as a core browser feature, it's hard to see how you justify dumping the description metadata into a third-party extension.
posted by ryanrs at 1:09 PM on September 25, 2018


Where are the descriptions in the html file? Is it the stuff between 'DD' and 'DT'?
posted by pjmoy at 1:57 PM on September 25, 2018


I think it's safe to say that messaging hundreds of millions of users for something that a tiny fraction of them use, with wording that leaves them fearful of data loss for something they don't understand and don't use, will erode more trust in Firefox than the alternatives.

I disagree that this is safe to say. What you're arguing is effectively that if a PR and user messaging problem is thorny enough, some users just have to accept data loss because Mozilla can't be bothered. That's a pretty good argument for never trusting a Mozilla project with your data, as well as advising anyone less technical you know that Mozilla's offering cannot be trusted.
posted by tocts at 2:26 PM on September 25, 2018 [8 favorites]


I've been holding out on upgrading to Quantum, hoping that the bulwark of defenses I've built up through a combination of hardcore privacy settings, security extensions, and 3rd-party anti-virus/anti-malware utilities will be adequate to compensate for a lack of security/bug updates while I cling to version 56 come Hell or high water

I did that for several months as well, largely for some of the legacy extensions (well mainly TabMixPlus, secondarily a few others) which were incompatible with Quantum. Eventually I became convinced of the security risk (not through any attack I was personally exposed to, but possibly in part because of comments on MeFi on the subject) and switched... to Waterfox, a Firefox fork which seems to be keeping pace with Firefox, and is generally compatible with Firefox extensions, including legacy extensions. My understanding is that Waterfox lags a few weeks behind Firefox in security patches, but that's acceptable to me.

(I don't use bookmark descriptions myself and don't know if Waterfox intends to remove them.)
posted by DevilsAdvocate at 2:38 PM on September 25, 2018 [1 favorite]


I don't even use bookmarks, I just write down Bing search strings on index cards and add them to an Agile wall. At the end of every sprint I ask myself "How well is this really working?"
posted by turbid dahlia at 2:58 PM on September 25, 2018 [4 favorites]


I have no particular dog in this fight, but a program should never delete my data without notification. Ever. Period. I would really have thought that the Mozilla team would know better than that, because it seems breathtakingly obvious to me and I am nobody's idea of a software developer.
posted by Anticipation Of A New Lover's Arrival, The at 3:07 PM on September 25, 2018 [6 favorites]


> Where are the descriptions in the html file? Is it the stuff between 'DD' and 'DT'?

It's the stuff after DD tags, yes. DT may not necessarily follow.

posted by Bangaioh at 3:55 PM on September 25, 2018


I am steadily proceeding through the Stages of Change Acceptance. So far, I have actually done and/or thought these things:
1. Confusion: "Wait, they're doing what? Really? Why?"
2. Googling: Brings up other, lesser message boards where people are wondering what's going on, and why. v62 release notes are not on the first page of Google results; you have to go to Help About, then go back from 62.0.2 to 62 to find them.
3. Anger: "THIS AFFECTS MY USE CASE, EVEN THOUGH NOT AS MUCH AS IT WOULD HAVE IN THE PAST! THEIR REASON IS STUPID! HOW DARE THEY! I"LL USE CHROME, MAYBE, OR LEARN ABOUT WATERFOX OR SOMETHING!"
4. Mitigation: Back my stuff up. Verify the information is still there, even though it's not really editable as easily.
5. Post to MetaFilter: "People should know! So they can back their information up and/or BE ANGRY!"
6. Bargaining: "So, for them to actually delete the data in v64, they'll have to go in and modify everyone's bookmarks databases. Surely a prudent process would back this database up before attempting to restructure it, right? I mean, it's one thing to remove data that you've hidden and that apparently most people don't use, but if something goes wrong with the database conversion and you didn't back it up, it'd be a shitstorm of epic proportions, right??"

Not at Acceptance yet, but they also haven't actually deleted people's data yet, either.
posted by Huffy Puffy at 4:01 PM on September 25, 2018 [1 favorite]


I'm trying out Waterfox because of the way Firefox broke TabMixPlus. I used a number of the features of that extension, but the big one for me was "open all links in new background tabs", which Firefox no longer seems able to do (even with an extension that claims to do it).

Worse, this breaks MF for me in fun ways. When I click on a link in a MF page that I've used the "x new comments, show" link to update, it opens the link in a new window, but then the back button does not go back to the previous state, but usually drops some random number of comments that have been loaded.

I'm also a bit leery of Firefox as I reported (18 years ago) that is still open (I get email updates every so often on it) and that caused data loss (though with the newer session saving features, it is less critical).
posted by Death and Gravity at 4:20 PM on September 25, 2018 [1 favorite]


I don't get what's so hard about this. It just needs to be another field where ever the bookmarks are stored.
posted by canoehead at 4:26 PM on September 25, 2018 [1 favorite]


I've been digging around the white papers on this trying to understand the background behind the problem, and this is what I've found.

1. Firefox user data is a bit of a mess right now with over 45 sources of information using 10 formats. So the long-term roadmap involves trying to simplify all of that and eliminate some bottlenecks.

2. Some of that data is stored in sqlite databases which have been a consistent performance problem. "It should go without saying that you should never execute SQL on the main thread." Part of this refactoring is to avoid that. The bookmarks editing interface appears to be OS-native, but I'm not sure how this influences the threading conversation.

3. The Places database records both bookmarks and history events. Since the description field is both variable-length and optional, it's normalized into its own table with its own metadata. This probably has consequences in terms of cache, temporary tables, and integrity checking. So I can see why they want to simplify the schema there.

There are some other advantages to moving that data to Web Storage (and one of the Web Extensions tutorials covers annotations of web pages). Web Storage gets a lot more performance and security scrutiny than Places. Having dealt with sqlite as a backend, the rationale that dropping a rarely-used table from Places can eliminate performance bottlenecks is entirely reasonable.

Mozilla is perpetually stuck in a no-win situation when it comes to Firefox. People complain that Firefox UI is sometimes slow, but also complain when slow, bad, and insecure legacy code is replaced.
posted by GenderNullPointerException at 5:17 PM on September 25, 2018 [6 favorites]


I gave up on Firefox when Quantum came out - I have a pre-Q version of Firefox with the plugins I use (like DownThemAll), and I now use Opera as my main browser.

Opera's bookmark system is horrible; it doesn't have a "view background image" option; I keep losing my plugins because I can't figure out how to find the settings. But it's a bit faster, and I have Firefox as the "open this when I need to download a swarm of things" backup browser, because apparently DownThemAll is impossible to replicate on anything else.
posted by ErisLordFreedom at 5:23 PM on September 25, 2018 [1 favorite]


People complain that Firefox UI is sometimes slow, but also complain when slow, bad, and insecure legacy code is replaced.

I don't think this is unique to Firefox; I think this is basically every legacy program ever.
posted by Anticipation Of A New Lover's Arrival, The at 5:59 PM on September 25, 2018 [1 favorite]


Zotero, which I've used in the past, now uses a standalone desktop app that connects to any browsers you use with an extension to bridge it. That's the sort of thing I can work with.

Anyway, backing up descriptions automatically and prompting users to save them in a format that whatever description extension gets created can import seems like the right way to go. Not sure if that's viable though.
posted by gryftir at 6:30 PM on September 25, 2018


I should probably make this an Ask, but since we've got a genuine Firefox expert here, I'm going to take advantage of it.

Lately, (I want to say for the past week or two), Firefox has had trouble connecting to websites. About half or a third of the time I open a new tab or a new window and type in an address, I'm met with the following message:

Hmm. We’re having trouble finding that site.

We can’t connect to the server at [site address].
If that address is correct, here are three other things you can try:

Try again later.
Check your network connection.
If you are connected but behind a firewall, check that Firefox has permission to access the Web.


If I refresh the tab once or twice, the site pops up.

I though it was just me, and that maybe I had too many tabs open, but I was just speaking with another Firefox user tonight, and he mentioned the exact same thing. He said he started noticing it about two major updates ago.

I'm on 62.0.2 (64-bit) for Windows, if it matters.

So is this an actual thing that people are experiencing due to recent build issues or something else more isolated?
posted by sardonyx at 8:11 PM on September 25, 2018 [2 favorites]


Having exact same problem, exact same platform & version...
posted by PhineasGage at 8:43 PM on September 25, 2018 [1 favorite]


jzb: "Look, if I use a program and store some data in it, I don't expect that program to delete that data ever without explicit notification. That is a violation of user expectations of the highest order. We're not just talking about a feature going missing, that's their data.
"

This. Firefox already has code to dump bookmarks; why not just run it at update time? Some one mentioned a security risk but presumably the data would still be in the profile same as always.
posted by Mitheral at 8:47 PM on September 25, 2018 [3 favorites]


The guy leading this change claims that with regard to this field it's impossible to distinguish user-input data from automatically filled data. At the same time he claims that the feature is used only by a very small number of people. I wonder how he figures the latter...
posted by patrick54 at 10:52 PM on September 25, 2018


right, but then you just? sign out of gmail? and you are no longer signed into chrome? i guess i just don't understand 01) staying signed into gmail all the time or 02) signing into chrome for anything other than one time bookmark syncs when you get a new computer? what are people doing this for.

Well, for me the issue is that I've never signed into Chrome (ie, for bmark sync, etc.) because I don't want to totally roll over to the data slurping asshole (and yes, I run my own mail server and calendar server). However, my employer uses Google enterprise apps and apparently forces updates/reboots without notice. So Monday I encounter the situation in which some personal data that I explicitly made an effort to not send to Google's servers might get slurped up.

Now, the real cherry on this fuck-you shit-sundae is that Chrome (at some point--don't know since I never had to use the "feature") added a "person" management page from which you can delete the "person" associated with the Chrome instance. So, I think, "oh, my data is already getting synced, I should remove this person from chrome to stop it." Nope (and here the quotes are person were a foreshadowing). Chrome apparently uses "person" to describe the local profile--bookmarks, history, saved sessions, etc--so, in a double UX fuck-you-anti-pattern... if you delete this "person" thinking you're removing the connection to Google server, you're actually deleting all your local data.

So...at least Mozilla didn't spray that tubgirl-stream all over its users. Yay?

I don't even use bookmarks, I just write down Bing search strings on index cards and add them to an Agile wall. At the end of every sprint I ask myself "How well is this really working?"
posted by turbid dahlia

JFC, give us a trigger warning....fucking ag*le bullshit...sprinting a g-damn double marathon

posted by MikeKD at 1:08 AM on September 26, 2018 [1 favorite]


the level of anguish that Firefox's users have been experiencing for the past few months. I hope Mozilla is aware that some of the abrupt changes associated with Quantum have been utterly catastrophic

I realize Metafilter is given to overwrought language, but this had me actually drop my mouth open. anguish. catastrophic. Really? I'm sorry your browser software is such an emotionally difficult thing for you. Perhaps you should consider using a different browser. Or stop using one entirely, or develop one to your liking.

FWIW, Quantum is fantastic for me. No anguish or catastrophe: only joy and excellence. It is solely responsible for me and several friends switching from Chrome to Firefox recently. I was about to smugly say Quantum has been a huge success in general but sadly the numbers don't back that view up; it looks like Mozilla is still in a long slow decline. That's a shame.


Back on topic, I had to scratch my head to remember what the Description field in a bookmark might be. It does seem a shame to throw that data away, hopefully either Mozilla will figure out some way to automatically export it for users or else a third party extension will preserve the data. It may be a tiny fraction of users who are affected but you can bet they'll be the noisy ones thinkfluencers.

I'm eager to hear about any improvements to the Places database. One thing I don't much like about Firefox is its browser history; the UI for accessing it is very limited and I feel like I often can't find pages I visited in it at all. The underlying database has the data I want but short of writing my own SQLite queries I can't get at it easily. That could be greatly improved.
posted by Nelson at 3:06 AM on September 26, 2018 [5 favorites]


TBH this just feels like of a kind with the "we redid the UI, and we also took away the functionality that will let you configure the browser to look the way you want" crap. I associate this sort of arrogant, user-contemptuous behavior with Chrome or MacOS; "we've decided that the way you use Firefox isn't the way we want you to, so we're taking away things that already existed and were important to you" still stings.
posted by Pope Guilty at 3:56 AM on September 26, 2018 [2 favorites]


> the level of anguish that Firefox's users have been experiencing for the past few months. I hope Mozilla is aware that some of the abrupt changes associated with Quantum have been utterly catastrophic

Also: "by abrupt changes" are you talking about the change to the extensions API which was telegraphed 2.5 years in advance of the old API being closed to give developers time to either rewrite their extensions or inform their users that the extension would no longer be available.
posted by urbanwhaleshark at 4:25 AM on September 26, 2018 [2 favorites]


I strongly suspect that if Mozilla hadn't ditched XUL, they would have become the next Internet Explorer in terms of security within five years.
posted by GenderNullPointerException at 5:40 AM on September 26, 2018 [2 favorites]


we redid the UI, and we also took away the functionality that will let you configure the browser to look the way you want
I know we love hyperbole here but you left out the parts about how XUL and XPCOM had a big maintenance cost and hindered major improvements: multi-process support, Servo, etc. all had substantial barriers supporting those customizations. This was all discussed in the open and the people who wanted to stay in the past weren’t exactly signing up in droves to help shoulder the significant cost of their demands.
posted by adamsc at 6:12 AM on September 26, 2018 [2 favorites]


I'm not upset that they changed the tech. Not at all! I'm upset that they a) changed the way the UI looks and b) in the replacement tech deliberately omitted the necessary bits to let us reconfigure the UI to look the way it used to.
posted by Pope Guilty at 6:19 AM on September 26, 2018


Informing all users of the deprecation of something that very few users use is a bad UX for the majority of them.

This speaks to one of the challenges we've got as an organization, when it comes to building new stuff and shipping valuable services to people while maintaining not just our hard, on-the-wire privacy and security promises but also the user's perception that we're upholding those promises. The fact that it's just about impossible for humans to observe software in operation occasionally makes life pretty difficult for us on this front, because it turns out that - in theory! - solving the problem you're describing looks pretty easy.

One hypothetical, unlikely-to-resemble-what-we're-doing approach would be, we change the client so that come next update, the code - now living and running only on your machine - peeks at your bookmarks, and says "hey, you've got something in this description field", and offers you a local remediation option of some sort. Only people who are affected by this issue see the prompt, nobody else is impacted. Problem solved, right?

This approach absolutely meets the letter of the law as far as our privacy standards are concerned, and on paper it looks great. Offers user more control over their data, totally private, nothing crosses the wire, excellent! But a chunk of our users will look at that and say "why is Mozilla looking at my bookmarks? Did I just (apologies, GenderNullPointerException) accidentally give Firefox my metafilter activity when I bookmarked that pizza place?"

And... we're not? We've put a bunch of extra work into making sure we're not? And that work might not matter, because perceived trust is only tenuously related to trust-as-implemented. Perceived trustworthiness matters as much or more, it's incredibly delicate, and getting it right while being as open and transparent about our processes as possible is, frankly, a lot of extra work in a context that constrains our actions in a lot of obvious and a surprising number of weirdly unexpected ways.

That's the job, though.
posted by mhoye at 8:30 AM on September 26, 2018 [5 favorites]


So is this an actual thing that people are experiencing due to recent build issues or something else more isolated?

I haven't heard of it, but I definitely want to know about it. Memail me?
posted by mhoye at 8:36 AM on September 26, 2018


Please update your browsers. I would go so far as to say that if you could only do one thing the ensure that your computer doesn't get screwed, it would be update your browser. Updating your browser is more valuable than all your anti-virus or any of your browser extensions to keep you safe.

I'd like to re-up this point - please keep your browser up to date. I think Firefox is pretty great, perhaps obviously, but I'd honestly rather you switch to any other major browser that you can keep current than rely on an outdated - particularly pre-Quantum - version of our product.
posted by mhoye at 8:46 AM on September 26, 2018 [5 favorites]


Perhaps you should consider using a different browser. Or stop using one entirely, or develop one to your liking.

Perhaps you should consider that not everyone uses a browser the way you do. For some people it's central to their job, their choice of browser is controlled by their company, and not being able to continue doing specific things in a company-mandated way is a real disruption.

My comment wasn't just about the bookmarks issue mentioned in the OP, but about a number of issues that I've found fervent discussions of in many places on the web. A lot of people are't happy that Mozilla has appeared to met users' complaints with "Ehh, you'll get used to it". I'm not a special-snowflake outlier crying in the wilderness.
posted by Greg_Ace at 9:07 AM on September 26, 2018 [4 favorites]


Yeah Chrome is basically a surveillance tool these days
[citation-needed] -- There's a lot of really bad shit happening on the web these days. Google is responsible for some of it. As far as I am aware, Chrome's behavior has always been fairly reasonable -- the recent update is far more innocuous than it's been made out to be (and, in fact, helps to reduce a potential source of user confusion that could cause you to inadvertently leak data to another user).

Baseless accusations like this one are not helpful. There's a lot of actual-surveillance going on.
I suppose it depends on what you consider "fairly reasonable". Here's an example: go into Chrome and
1. clear all your browsing data. Literally click "Clear browsing data" and pick the time range "All time."
2. Go look at what browsing data it didn't clear: go into Content Settings -> Cookies -> See all cookies and site data. Huh. A whole buncha google.com cookies (including your advertising SID and NID) and localStorage and your doubleclick.net advertising ID cookie. Huh.

There's literally no reason to keep that doubleclick.net cookie around except to track you for ads. Even if you try to give them the benefit of the doubt along the lines of "we don't want the user to be inconvenienced by getting logged out of gmail", the only reason to include doubleclick (and, lets be honest, the root google.com ad cookies), it's Google saying, "we make this browser, so we're going to clear everything except those we profit from."

Which, hey, if the Chrome folks want to redefine what clearing your browsing data means, that's certainly something they can try. They even recently put a little text disclaimer on the "clear browsing data" dialog now that says something like "won't log you out of your google accounts," but I'd be real surprised if many people understand that to mean "we won't delete any google ad tracking cookies or localStorage."

This is not the only super-fucking-shady thing that Chrome does.
posted by introp at 12:37 PM on September 26, 2018 [4 favorites]


Does Chrome do the same "we don't clear your data haha!" thing to European users? Because considering the GDPR, that seems like a really bad idea.
posted by Nelson at 2:03 PM on September 26, 2018


It looks like there is no way in current Firefox to see all my browser cookies...?
posted by PhineasGage at 2:34 PM on September 26, 2018


It looks like there is no way in current Firefox to see all my browser cookies...?

Got to Preferences, then Privacy and Security, and under Cookies and Site Data is "Manage Data".

I also work at Mozilla. mhoye is amazing
posted by potch at 2:42 PM on September 26, 2018 [1 favorite]


While we're talking security, it looks like Firefox won't store passwords for insecure connections. I don't understand the rationale (since I'll be typing that field in manually anyway) but I guess maybe? However, "insecure connections" includes my router, which has a very long and secure password that I have to enter manually every time. As far as I can tell from Googling there's nothing I can do about this, except maybe find a different router that accepts https connections.
posted by Joe in Australia at 6:16 PM on September 26, 2018


users are in fact entitled to control over their own data is a real thing we believe in

mhoye gave an inspiring talk about this to GTALUG last night. I'm sure it'll hit their Youtube feed soon, with luck complete with the esr and cryptocurrency jokes too.
posted by scruss at 10:29 AM on October 10, 2018


« Older My favourite too   |   Consent should be mandatory Newer »


This thread has been archived and is closed to new comments