I find it interesting that the gif shows the Tree Style Tab extension without the default Firefox tab bar, but at least in FF57, the tab bar can only be hidden with a custom userchrome.css, so the otu of the box experience is flawed at the moment.
This is at the moment my biggest gripe with the extension, and I think the interview / article could have shown the default state of the extension, and pointed out that an API for hiding the tab bar is at least being worked on: https://bugzilla.mozilla.org/show_bug.cgi?id=1332447
Thanks for pointing this out, I got excited for a moment thinking they'd finally fixed this. The article seems pretty misleading by leaving this detail out: using this extension shows you two sets of tabs which is kind of ridiculous.
Yes, there's this problem too if you run Tridactyl (Vimperator/Pentadactyl's successor).
You may want to have no address or tab bar given that you enter URLs in a Tridactyl's own field, and you manage tabs as buffers. Even worse, Firefox own address bar steals focus in some scenarios.
Does Tridactyl suck as much as Vimium-FF on FF57? I would guess so. I used to run VimFX, but I'm out of luck now. It seems we have to wait for the DOM to load before we can use our keyboard shortcuts...
I just tried the userChrome.css hack and it breaks a lot of stuff. Tree Style Tab stops being able to tell which tab is active, my address bar only affects the first tab, and my window buttons are all blank. Looks like I'll have to wait for the tab hiding API to try this out.
/* to hide the native top tabs */
#TabsToolbar * {
visibility: collapse;
}
/* to give the top toolbar a bit of height so the window controls don't dangle off to the right */
#TabsToolbar:before {
content: '';
display: block;
width: 100%;
height: 20px;
}
/* to hide the sidebar header */
#sidebar-header {
visibility: collapse;
}
Have you tried the built-in titlebar setting in Customize? There's an option at the bottom to turn on the regular system titlebar so that it displays as a normal window.
I don't understand why they don't just create an API that allows extensions to modify userChrome.css. It seems like this would solve a lot of the complaints people have about the new extension framework, and it would be a way better experience for non-tech-savvy users.
Probably because any change to the Firefox UI would potentially break an extension using such an API, and part of the point of switching to WebExtensions is to have extensions that never break.
Also, such an API would allow an extension to make the browser UI unusable, which would make figuring out the problem and removing the offending extension quite hard.
Mozilla basically titled the security vs. versatility balance massively in the direction of security with no sudo-equivalent for AMO-distributed addons.
because they're inconsiderate of Tree Style Tab users. They could really well include a setting to disable the tab bar on top, but they're not doing it.
There are no security reasons for not having such a setting. None.
There's a major problem with addons doing stuff the users don't want, and the tab strip is a critical portion of the user interface so "just let them hide it" has problems with abuse. So it needs to be something that can still be autoreviewed, and it needs to be something that gives clear indications to the user as to what is going on.
They work on it. The main problem is simply that Mozilla fucked up the transition from legacy addons to the new system. There are so many APIs and functionallity still missing, that most commercial companys would have been killed over it. But with OpenSource this has become kind of normal to accept it. Just Remember KDE 3 -> 4. Python 2 -> 3. Gnome 1 -> 2 -> 3. Breaking things is just to common now.
It was the first extension I installed when I switched from Chrome to Firefox 57 (Nightly, at the time). Then I used the userstyle css document to remove the tab bar at the top. As a non-web developer I've been extremely pleased and haven't had any desire to switch back. I also prefer Vimium over cVim, the Chrome extension I was using for vim keybindings.
The developers of Tree Style Tab have really done a great job.
Vimium is available on Chrome, in case anyone gets confused.
I got a nightly build a while back but didn't feel impressed enough to change my $BROWSER. Maybe now I'll give it another shot, since TST w/o a top tab bar sounds fun.
No, it's impossible for the moment to deliver reliable keybinding-extensions. Firefox has no API for it, so all solutions need to work by injecting a script into each website, and grap keys from this context. And that simply sucks and leads to many problems. They are supposed to work on a solutions, but it likely still needs 6+ months.
Besides that, vimium works well enough under perfect conditions, and is quite mature as it is a port from Chrome. Tridactyl is also promising replacement for Vimperator, but still needs time to grow, because they just started some weeks ago.
I did this exact same thing but did not see an easy way to reassign the F1 shortcut to toggle the tree style tabs at a quick glance.
I looked through the source code (manifest.json) where it was defined but didn't go looking for how to have the changes seen by firefox, if I were to just change it.
I didn't remove them but I moved them back to where they make the most sense with this:
$profile_dir/chrome/userChrome.css
/* Tab bar below Navigation & Bookmarks Toolbars For best results, show Title Bar or Menu Bar */
#nav-bar { /* main toolbar */ -moz-box-ordinal-group: 1 !important; border-top-width: 0 !important; }
#PersonalToolbar { /* bookmarks toolbar */ -moz-box-ordinal-group: 2 !important; }
#TabsToolbar { /* tab bar */ -moz-box-ordinal-group: 3 !important; }
I also prefer the old style menus so I re-enabled them with View->Toolbars->Menu Bar.
Another quick tip: those big ugly spaces between the buttons and the address bar can be removed with a simple right click->remove on the empty space.
No idea if this is intenional or if it was just for the drag space / design, but what it also does, is that it moves the start of the URL more into the middle of the viewing area.
And the start of the URL contains security-critical information (http/-s and domain name), so having that closer to your eyeballs is generally helpful, even if it means cutting some of the URL off at the end. Especially so for average users which is ultimately the user group that's mainly going to be using the default.
Since on most platforms the tabs go all the way to top of the window now, there isn't room to grab the window to move it around. Those spaces act as extra drag targets.
To hide the tabs on top and the sidebar title only for the Tree Style Tab extension, use this userChrome.css
/* Hide horizontal tabs at the top of the window #1349 */
#TabsToolbar {
visibility: collapse !important;
}
/* For only Tree Style Tab sidebar #1397 */
#sidebar-box[sidebarcommand="treestyletab_piro_sakura_ne_jp-sidebar-action"] #sidebar-header {
display: none;
}
There are two reasons, the first is there. The second reason is my job. I'm employed as a support engineer for company-use of Firefox/Thunderbird for 11-12 years, and the knowledge and trust are based on my addon development experience. So, I think I already monetized from addons and donation looks like a double-dealing for me now. (So after I retired I may accept donation, if Firefox and my addons still alive.) Anyway, I just say thanks to all!
> i hope to keep having a privilege to say "no" about requests not matched to my vision
I totally understand and respect that. I bet it leads to better, more focused software rather than 10 mostly-similar projects that all almost do what you want. Kodus to him!
100% agree, I held out for so long before being forced to upgrade my Firefox because losing multi-row was like cutting off a leg. I don't know why there isn't a simple checkbox to make tabs create new rows, if I don't like the loss of space, I'll close tabs, but now I can't seen what tabs are open because of these dumb left-right arrows and no space! I miss Tab Mix Plus!
You can use this as a base for your own multi-row-tabs experiments: https://www.reddit.com/r/FirefoxCSS/comments/7dclp7/multirow... I use a modified version of this and I am kinda happy. With a real extension it could be much better but for now at least it's better than the standard tab bar.
If anyone has a better guide/config: Please post it here!
There are other tab behvaiors that TabMixPlus offers including going to the last used tab on close, open in foreground tab in the context menu for links, scroll wheel to cycle through tabs, and some other things I take for granted and then get disoriented when I use FF57.
Without those, it's painful. I will stick to the older Firefox (or maybe an alternative like Waterfox) until it becomes impossible to use new sites with it, or if somebody (myself included) reimplements these features in a way compatible with FF57.
This is the single most infuriating thing for me and I just don't understand why this is not a standard option, especially since it works perfectly fine by default on Chrome!
Probably because it's used for scrolling the tab strip when the amount of tabs is so high that their width stops shrinking and further overflowing tabs are clipped from the strip area.
I wasn't even aware of multi-row tabs because I was still using Chrome, which dynamically resizes the tabs until they just show the icon, which usually is sufficient to know which tab is what. Now, in Quantum the minimum tab size is HUGE in comparison, showing the first ~5 letters. Instead of resizing smaller, I have to use click the arrows left/right which change from light gray to white (barely noticeable) when tabs are hidden. I've been using Quantum daily since the release and I'm really fighting to not switch back to Chrome (and don't even get me started about the LastPass extension bugs/missing features).
What is the use case for multi-row over tree style? I feel like tree-style allows me to easily see all my open tabs and their hierarchy. With multi-row your tabs are still scrunched into a small area that does not allow you to read their title or easily differentiate them when you have many open.
I'm surprised there hasn't been more backlash about Quantum regarding specifically this. Tab Mix Plus doesn't work, so Quantum lacks the killer feature of multi-row tabs. And all this nonsense about userchrome.css as a workaround is just .. no. I am not taking on the burden of maintaining that thru future versions.
I'm stuck at 56.02 and that's for sure not great, and if I roll back to ESR I lose my history. And staying stuck at 56 is so far impossible, something keeps auto updating it even though I turned the update off in config. I've resigned to just reinstalling 56.02 periodically.
For the first time in forever I'm seriously considering ditching Moz. /rant
> I'm surprised there hasn't been more backlash about Quantum regarding specifically this. Tab Mix Plus doesn't work, so Quantum lacks the killer feature of multi-row tabs.
Hardly anyone actually used this or even knew about it, that's why there's essentially no backlash. (Which is to say thousands of people used this, but that's just not many at all at the scale of Firefox.)
Also waiting for a TabMixPlus replacement. I've updated FireFox on a few of my other systems as experiments and used userchrome.css to tinker but my main development machine will stay at v56 until I can properly replicate my workflow in 57+
As much as browser tabs are a broken band-aid,[1] TST is a vital and tremendous improvement over flat horizontal tabs, and absolute clobbers Chrome for productivity. With the performance improvements of Firefox 57, there's absolutely no question which browser is superior.
The wonderful thing about TST is that they create an inherent structure to ones tabs: those opened from the same root create a subtree, can be expanded or collapsed collectively, or moved, or deleted, as a group.
The power and flexibility this gives over Chrome's abysmal tab management is simply staggering. For whatever the reasons, Google has been pursuing the minimum viable user,[2] with highly predictable results: advanced users are crippled through a brutally insufficient interface. My experience on numerous bugs and issues filed against Chrome confirms that Google have no interest in addressing this.[3]
That said: I'd really like to see a description of the "Container Tab" concept, as I can't make heads or tails of what this is supposed to be, do, or be used for.
Otherwise, congrats to the Firefox team. I'm enjoying the new features and capabilities, and the mainstreaming of TST.
________________________________
Notes:
1. A view I apparently share with Adam Stiles who invented the things, though his comments on same were on such a clickbait-laden site I never bothered linking it.
3. I've launched a Denial of Feedback attack on Google for its various products. I no longer file bugs or report issues. I encourage others to join in this attack.
> For whatever the reasons, Google has been pursuing the minimum viable user
Probably because this covers 95% of the total user base. Most people are not developers and only get confused by extra features. In fact, most people I've seen using the browser have no idea what browser is and find even basic tab interface complicated.
The further argument of why this is a problem is given in the linked article.
TL;DR: the minimum viable user has exceedingly poor computer skills. And even this, based on a large multinational study, is only about half the population. "Advanced" is still fairly rudimentary, and comprises about 3-8% of populations in countries studied.
If it's not possible to bridge that divide, the situation bodes poorly.
(And the dynamic shows up in numerous other places, though I'm not sure that the capabilities divide is always so wide.)
You know why there's lists of ingredients on all food, and all sorts of other info in manuals and elsewhere even thought most people never read it? I know it's partly because of regulations, but ideally, it would be because of pride of craftsmanship on behalf of the people making them. And at any rate, if someone doesn't want to know what they're eating or using, that's their problem, but they don't get to dictate that to anyone else no matter how many of them there may be.
But in this super new fangled world, where less is more to the point that worse becomes better, there's hardly time to write documentation because you have to shovel more product. Why worry about things like link rot that actual webmasters (that would deserve a title implying mastery) of olden times discussed, when you can have nothing there to begin with, right?
Decent adults have handled and will forever handle everything as follows: You try to enable autonomy and creativity and knowledge in those who have the means and drive for it, and try to make it so it doesn't get in the way of the rest too much. Nothing against elegance and simplicity, but not at the expense of moral and intellectual decency. And what's so impossibly hard about having "advanced options", or even several builds? That users will drop the product for the competition that doesn't even exist anymore? People can magically learn to drive cars without becoming murderers ever when they turn 18 -- but we cannot RISK them getting confused with options in a browser. Give me a break. This is intellectual dishonesty contortion news so much of the time. Hacker is NOT supposed to be the superlative of hack, damnit.
Google is an ad company. Ads are for making people act against their interests, not for informing them. That's where the journey is going for those onboard with it.
> In fact, most people I've seen using the browser have no idea what browser is and find even basic tab interface complicated.
That doesn't mean extra features confuse people necessarily. Also, "most" isn't 95%, nice try there.
People, including developers, tend to ignore the things they don't understand or haven't had time to look into yet. When I open about:config I would have to look up most of that stuff to know what it means, and even then I might leave most of it on default. But I can still go down the list and see if something looks useful to me, and learn more about it if I want to. When I read a manual that has multiple languages, I don't read it every page with a confused stare until I get to the language I actually know. I just skip to the bit I need.
Most kids can't read and write when they get born. Why teach them? Why even run the risk they might learn it from scratch by leaving books around? If this is the information age (heh), you're with the priests of mass illiteracy. Google determines what you see, not your search query. Facebook organizes your content and what you see of what you're interested in, not you. We went from personal computers to walled garden spy devices, and saying that's what people want, as if many in the "industry" (I put it in quotes because that comes from industrious, and I'm not seeing the diligence and craftsmanship that would warrant such associations) don't want that to the point of pushing and pushing and pushing to get it, is essentially victim blaming.
(sorry for continually adding to this post, but you know how game consoles where supposed to be superior because you don't have to fuzz about, it just works? now they have all the downsides of computers but none of the advantages, they actually start leaking into and degrading PC gaming, and the people who said "fuck that shit" in the 90s were proven right, once again)
And it's infected Firefox, too. E.g. open the options, and you have to click on "Firefox will [Remember history]" to see there are custom settings available. The only way for a newbie who heard of cookies to find the cookie settings is to click on everything. Lame, especially considering the padding and line spacing. It's not users who are confused, it's developers who are lazy. They just assume because this is the direction "everybody" is going in, any shoddy rationalization, and fake statistics like "95%" and "most", will do. After all, it wasn't your idea, you're just going along.
I'm currently trying out FF57 with the TreeStyleTab plugin and css fix. I switched to Vivaldi a while back, mainly due to performance issues.
Vivaldi was faster, but has a lot odd bugs, it's not OSS, and has no real public bug tracker.
After some recent UI lockups, I decided to try FF57 again. Most of the extensions I use have been ported, but after just a day or two, I started to hit real performance issues again. I'm not ready to give up on FF yet, but I do feel like many of the performance improvements might have been overrated.
What would help is to activate the Gecko profiler[1] when the performance issues occur. The cause of these performance issues might not be in Firefox itself, and the Gecko profiler would allow to find out where is the issue.
If you're looking for performance, don't go with Vivaldi, at least if you use many tabs.
With many tabs (>= 100 or so), it becomes really sluggish. For example, opening a link in a new tab can take approximately five seconds. Chrome, which it is based on, does not have this issue. Chrome is however unusable (with many tabs) for other reasons (lack of lazy loaded tabs essentially causing a DoS for many minutes when starting it, at least on Windows).
Shameless plug: try https://cretz.github.io/doogie/ which is my Chromium-based browser with trees. It is not the most mature thing out there, but I use it for most things. But geez, that process-per-tab of Chromium approach can really blow up your memory.
uBlock Origin, TreeStyleTabs, Reddit Enhancement Suit (I should just disable that; stopped using Reddit after their CEO edited comments and their warrant canary got removed), Open In Browser and Mega.
Only thee of those are active on more than one website. If Web Extensions are suppose to be this great improvement for security and performance, extensions themselves shouldn't matter, or should at least show up in about:performance if there are issues.
Despite being one of the best adblockers, uBO is a small but noticeable performance hit to me- it's still worth it though. Nature of the beast I suspect.
Can you be more specific about "noticeable performance hit"? A whole lot of work has been done to avoid "noticeable performance hit", so I would be curious about a reproducible case for this to find out the cause.
I will attempt to come up with a test case. Mostly I just noticed when resetting Firefox or with a new install, "hey this is snappier than usual" until I install uBO. Not sure if it's resources, UI responsiveness, or rendering speed. uBO is awesome, and I appreciate all the work.
I thought you had hard data backing up your claim. Frankly I am quite skeptical at any unspecific "noticeable performance hit" claim. uBO's overhead is typically in the low milliseconds range, which amount to noise-level overhead really when compared to everything else occurring at page download/rendering time.
Well, I went straight to Facebook as it's got a lot of ads, and disabling uBO does appear to speed up the "long tail" load of FB.
I'm not complaining, and if you're confident in your overhead times feel free to disregard. Just unscientific observations. I still love uBO and use it across the board.
I did the same thing. I gave FF about a week before I got frustrated and went back to Chrome. FF is still noticeably slower in terms of basic UI actions (opening/closing tabs, menus, etc.) and every hour or so it would start hammering my CPUs. I don't know why.
Have you looked at about:performance and also when you hit this, then create a perf profile using https://perf-html.io/ and submit it as a bug to bugzilla.mozilla.org. If possible do both using Nightly. In my experience the developers are very(!) welcome to good bug reports.
Upgraded to Firefox 57, many extensions that work before didn't anymore, especially the Session Manager and a little extension that cleared cookies of the current website, which is badly needed for web development. After waiting for a while and nothing seem to come up, I've finally broken down and started writing extensions. The new WebExtensions api is not so bad. I was able to whip up a prototype after a day. The API has restriction due to the security model but there are ways to get around to get the job done.
I use Tab Session Manager as a replacement but it's clearly not the same.
In fact all my extension replacements are inferior to their previous counterpart (not even considering having to edit an obscure CSS to customize the UI).
I still have to restart the browser regularly but now I can't have a restart entry in the file menu. Tip: shift+F2, "restart".
My strategy has been to duplicate everything in Chrome and use both, with pinned tabs running on Chrome, especially for Google apps.
Another tip: For some reasons E10s wouldn't activate anymore with the update to FF57 (check it with about:support) and I had to create a new profile to make it work .
I want an one-click button on the toolbar to clear the site cookies. Plus it's a small enough scope project to let me try out the WebExtensions development process.
It creates automatic snapshots of your sessions. If your session gets wiped or corrupted, you can fall back to one of the previous sessions. Or if you just want to go back to the tabs you had opened at a certain point in time, that's possible too. All of this is important for those of us who keep lots of tabs open. I personally treat my open tabs more like chronologically organized bookmarks and usually have over a thousand open.
I went back to Firefox for Quantum after a very long hiatus — mainly Chrome then Opera (Blink) — and TST was a main factor of my switching. Once it's customized to my liking, it's even better:
Since the most recent version of TST, the author has added a barebones style, which can be customized at-will thanks to the custom styles textarea in the addon's settings. I've fired up the Browser Toolbox to get all the correct CSS selectors and then customized things to my liking. I've shared a gist on the addon's github issues with a version of my custom styles. I've since added a few things. I could create a new gist if you are interested
Looks nice, quite similar to my setup. But you seems to also be using a number of customization that aren't possible with the extension alone - curious how you configured the auto-expand on mouseover?
Indeed! The auto-expand bit is largely done through userChrome.css (someone shared a solution that I've tweaked very lightly, props to the original sharer)
So basically he recreated something that already existed because Firefox broke their userspace and he was forced to. The fact that Mozilla advertises this as progress is just weird.
Perhaps I'm just grumpy that I upgraded Firefox and my favorite extensions broke. There was no warning, they just broke because Firefox broke their API. And I mean I get it, I actually wrote a rather complicated plugin for Firefox using WebCrypto.
Firefox's userspace is webpages, not extensions. Extensions before 57 are like out-of-tree kernel modules. Extensions after 57 are a lot more like userspace, though!
>Extensions before 57 are like out-of-tree kernel modules
Its not that you shouldn't compare firefox api to unix userspace, but that firefox's api equivalent in unix is not userspace, but out of tree extensions.
In which case firefox isn't at fault for breaking api promises it never made; extension authors are at fault for not keeping pace
And firefox never promised you, the user, that the extensions would work in all versions of firefox; rather the extension author (may have) promised that, when they shouldn't have (because they weren't in a position to uphold it).
If the firefox api isnt the equivalent of unix environment, then firefox isn't wrong to break extensions. Its just part of the natural lifecycle of the application, unless firefox changes that stance, and does make extensions part of their promise.
Where are they advertising anything as progress in this interview? They're just trying to share experiences people had with WebExtensions, so others know what to expect and for Mozilla devs to know where things can still be improved.
Only on nightly. In the production-version they activated it in v56, the version before the break. And as I remember there were no real information for the casual user what it means, or what they should do. Not even a choice when autoupdate hit.
As I wrote in the article, indeed I worried about death of addons in last 2 years. But, from various viewpoints I analyzed Mozilla's statements around addons, and finally I decided to forget past and shift my mind moving toward to make things better. So the story of the interview is almost correct.
Thanks for all your work. If you had a donation option, I'd be happy to limit my contribution to what I've gained from the past ~10 years of usage with literally no expectation of continued development.
I really appreciate you writing up and sharing your migration story and raw opinions, as well as all the great code you've written. Thanks you!
I've done some xulrunner app development as well as IE ActiveX control browser extensions in the past, including various implementations of pie menus, and I would like to learn more about the latest Firefox, Chrome and Safari extensions APIs, so I can implement pie menus properly for those platforms.
That's why I'm so happy to be able to read your migration story and raw opinions! Reading about you experience is even more valuable than reading the source code itself! Thank you again! ;)
I've implemented pie menus in jQuery for web pages [1] [2], but they really need to be rendered with transparency in their own arbitrarily shaped popup overlay windows, so they're not confined to the browser window, like I was able to do with ActiveX controls in C++/MFC and also X11/TCL/Tk back in the 90's [3], and with NeWS in PostScript back in the 80's [4].
I noticed that Safari extensions have an API for creating popover windows called SafariExtensionPopover [5] [6] [7]. But they must be rectangular, and the documentation doesn't mention how their position is defined (but I presume they appear beneath the cursor or invoking toolbar item). So there doesn't seem to be an obvious way to position or shape them. (Could you get the popover's window, and set its position pixel-perfectly regardless of borders?)
As far as I can tell, Safari's popovers are only good for traditional linear menus that pop up below the cursor or toolbar item, because you can't even control their position to center them on the cursor like a pie menu requires.
Do you know if there any way to create a pop-up window in any position or size, and even to set the background color transparent, so I can make arbitrarily shaped and popup windows? Even rectangular overlapping popup windows would be better than having to clip them to the window frame.
Do you have any suggestions for how I should approach the problem, please? What kind of browser extension API do you think could be designed or modified to support arbitrarily shaped pop-up windows for pie menus and other techniques (like tabs sticking out the edge of the window [9], for example)?
(The X-Windows SHAPES extension [10] is a terrible example of the worst possible way of doing that! ;) )
You can use native messaging to send a message to a script or executable that crates that arbitrary and transparent window. Nothing inside webext allows you to do exactly what you are looking for though from inside the webext.
Before the WX version, we were able to use a GUI to choose the style for the fonts in the tabs (or maybe that was in Tab Mix Plus, which hasn't been ported?). We no longer have that feature.
I was finding that it was hard for me to pick out which tab was active, but found that the extensions options provides a mini-editor for giving CSS to customize its rendering. OK, it wasn't hard to find that - it's just at the bottom of the Options window. But I thought I'd offer my settings, where I render the active tab larger and in a bigger font, and some other changes make it more usable for me. Maybe that will save somebody a few minutes in figuring it out themselves.
Most likely it was Tab Mix Plus. Nothing in Firefox lets you choose individual fonts for UI components. It's all or nothing: either content or whole UI.
Everyone should give it a try. I did give it a try. Before Firefox 57 I used Tab Groups to organise my tabs. I always wondered what the fuss was about Tree Style Tabs. Now I know, it is a wonderful addon.
When I'm working on a specific problem I tend to open a lot of tabs, so my workflow was to create a new group, open all the tabs I needed while I worked (sometimes I'd open the same link in several tabs because it was easier than trying to find where it was) and, at the end, move whatever tabs I wanted to keep (for reference or future use) and closed the whole group.
That and good names for groups made easy to find anything even with hundreds of tabs. If something else came up and I had to drop everything I could open the groups view, select the appropriate group and have an almost-instant context switch. Then I could go right back to whatever I was doing and pick it up with the same ease.
With Tree Tabs but without Groups I can do something similar by using the root of each subtree as a group. However, it's not as nice:
- You can't change the title of the tab, all you can do is use a website whose title describes as best as possible the contents of the subtree, which makes groups like 'email' and 'Hacker news' easy to recognize, but complex tasks almost impossible (not to speak of two similar but separate tasks)
- Even if I close all the subtrees and only leave open the one I'm working on, the titles are still there, always visible. I find that it makes it more difficult for me to concentrate (and increases the chances of procrastination)
What do you mean "environments"? Do you mean "containers"? The ones called "personal", "shopping", etc?
I've done a quick search for environments in case there's something I missed, and found nothing, so maybe it's a translation issue?
If you DO mean containers, then yes, it's a wonderful idea and I use them extensively[0], but has nothing to do with Tab Groups. If you boil it down, Tab Groups was just an UI extension, all it gave us was another view of all the opened tabs and, if you ask me, that idea matches perfectly with containers; just make each group a container.
My only issue with containers are that you can't move one tab from one container to another. I understand why and the complexities (and maybe even dangers) of doing so, but strictly from an user perspective it feels... wrong. It feels like an arbitrary limitation. Of course, you can copy the url, open a new tab in the container and paste it there, but then you lose all the history (and might need to log in again).
[0] With Tree Tabs I've made every root tab a different container and every time I open a new tab from that root, I made it from the same container.
It is :-) German Firefoxes translate the containers as "environments" which is somewhat logical, at least. I just can't really manage to integrate them into most of my workflows since they basically void my cookies.
I wish there was a way to collapse the tab war only when the tree style tabs are visible. I find the extension very useful when I have a workflow that requires categorizing many tabs open, but the bar uses too much horizontal real estate to justify keeping it open constantly.
Click Help -> Troubleshooting information (or type about:support into the address bar) and it will show your current Profile Folder. It's in the chrome subfolder under that.
For me it's at ~/.mozilla/firefox-trunk/047mvdpd.default/chrome/userChrome.css (your profile id will be different, and it'll be firefox instead of firefox-trunk if you're not running nightly).
Wow thanks, the directory was absent in my profile, and I was thinking maybe my version doesn't do this. It's all nice and (slightly more) spacious now!
But, with WebExtensions, authors can just say, “It is impossible due to the API limitations.” I think this is a very helpful and important change in WebExtensions.
That was a punch in the gut. Now the intentional gimped-ness of the new API is being held up as a positive?
The interview article is mainly for addon developers, this is the reason of the answer for the question. In past days I was a "people pleaser" and I didn't say "no" from my frailty, then my old addon became too fat, and finally dead. I hope that addon authors publish long-living small addons instead of short-live too much addons.
The sentence in the interview means that now there is a boundary of responsibility between addon and Firefox. If a feature request is impossible to be implemented due to API restrictions, it have to be solved by Firefox itself. Thus if I have strong motivation to implement a new feature requiring new API, I still file a proposal of new API, and I think other addon authors also should do it based on his demand.
Interesting setup. Thoughts on it compared to TST: Appreciate that this method is built-in. Very much like the automatic Closed Tabs group. Dislike that it disables middle-click to close a tab. Dislike that there's only two hierarchy levels. How do you make a new tab group without the tab bar, though? Also greatly dislike the added step of manually making a group compared to it being automatic with middle-click, even though it's a logical consequence of the rest of the design.
Have they added syncing yet? I tried it months ago and liked the browser fine, but with how many devices I use it's just not feasible to do bookmarks locally in 2017.
I wrote my own self-hosted cross-browser Xmarks alternative which happens to work on Vivaldi but is currently only tested on Firefox. Turns out cross-browser WebExtensions are not quite as easy as I thought. Alpha-state code and website: https://www.ymarks.org
Q: On a scale of 1 to 10, with 1 being the easiest, how difficult was it to write to the WebExtensions APIs?
In general, it was a 2 or 3. WebExtensions APIs are simple enough and clear for extensions that just add a new button, sidebar, or other unique feature to Firefox. However, there are some undocumented behaviors on edge cases, and they might be closely tied to a specific Firefox release you’re working on. If you’re writing extensions that change Firefox’s behavior, you will need to dig into those undocumented behaviors.
Sounds like the same pitfalls as writing a XUL extension...
By default it did not retain tab groups after restart. Is this as intended? I uninstalled it because of this.
I have my hopes in Simplified Tab Groups [0]. It was also faster on my old ThinkPad T42p, probably because of the lack of animations. However it's not yet ported to Web Extensions
This extension makes me sad. Not that this extension isn't done well, but I think the existence of it, and that it's on top of HN reflects a sad state of UI in our browsers.
Let talk about tabs. The reason that tabs exists is because we don't want to switching between windows when we have multiple tabs open. But think about it, why are tabs any better? When I have about a dozen tabs opened on Chrome, all I can see is a tiny favicon. Any more than that I only see an (x). Firefox and Safari are a little better in that their tab bar scrolls, but still only slightly.
The reasons we open multiple tabs is primarily for cross-referencing, and secondarily "preview", i.e. the intermediary state of article quality evaluation, the state between I-dont-want-to-open-this and I-want-to-save-this-for-reading-later, and/or I-like-it-so-much-I-want-to-bookmark-this-for-eternity.
How are tabs, tab ordering, top sites, tab preview thumbnails, tree tabs and any of these incremental "improvements" on tabs going to help us do xref and eval articles?
For cross referencing, I typically open one "root" page, and then drill down to the rest. For this use case, a hierarchical arrangement of pages is probably a good idea. An MRU and/or stack on top of that is even better because maintaining a tree traversal ordering in the head all the time is hard. Too bad FF took this one feature that doesn't scramble my thinking process away years ago. I know there's control tab, but that's not good enough because that's a "cross hierarchy" mechanism. As soon as I click on a tab that belongs to different "group", MRU is messed up. I really want MRU by groups.
To address this, I need to address why I'd click on a different tab in the middle of xrefing a group of tabs. This is usually because I'm taking a break, so I press Cmd-T and introduce a new group. The problem with how tab orderings are typically done now is, what the break is over, I still have a bunch of "pages" opened waiting to be evaled, but I do want to get back to the previous grouping. How do I do that now? My tab bar is full of tabs in various different states now. To solve this problem, I can open a new window and put a new group there, but the problem is, the screen is not that large a laptop, and I frequently maximize my browser. So any extra windows are just going to be hidden in the back and I'll forget about them until a month later.
I don't really need an actual side bar that takes up screen space all the time to show me the hierarchical relationships of tabs. I mean, it's only useful if MRU doesn't do what I want.
I don't know if it's just me, but I feel like the 40 year-old GUI pattern just doesn't work for browsers anymore.
Vertical tabs in tree structure is currently the best way to use web, especially when you pair it with auto-unloading of inactive tabs. I'm also disappointed that nothing better appeared in last few years. Developers seem to be extremely conservative about UI. With WASM it seems that browsers will take over most of OS functionality and the UI is getting in the way.
IMO we should ditch tabs, bookmarks and history. They all implement the same thing - a way to reach a place you've visited before. And they all require to be micromanaged in specific ways. The browser already knows which sites I've visited recently, and all the backtracking and branching I did. This information could be visualized as tree so that I can backtrack to any node and resume browsing on a new branch. The browser should also be doing memory management so that revisiting recent site is as fast as switching to another tab. And while we are at it, why not allow full text search across recently visited sites?
"the best" is an exageration. The web is named so for a reason: it's a graph, not strictly hierarchical. Which is rather difficult to represent in a list format.
For wikipedia e.g. I really want to have a graph instead of a linear history view, in at least 2 dimensions: ordered by time of visit and connection to related articles. I'm kinda ashamed I never got around to make this in 10 years.
> MO we should ditch tabs, bookmarks and history. They all implement the same thing - a way to reach a place you've visited before. And they all require to be micromanaged in specific ways. The browser already knows which sites I've visited recently, and all the backtracking and branching I did. This information could be visualized as tree so that I can backtrack to any node and resume browsing on a new branch. The browser should also be doing memory management so that revisiting recent site is as fast as switching to another tab. And while we are at it, why not allow full text search across recently visited sites?
Yes! This! Essentially, something like Emacs's undo-tree, but representing the tree in a more compact form, using some variant of D3 or Graphviz's tree layout algorithm.
I use tabs as a todo list. For example on HN, I quickly open all the comment pages of headlines that look interesting. Tree-style tabs are excellent for this use case, especially because I can collapse the tree when I need to move onto another task.
> I don't know if it's just me, but I feel like the 40 year-old GUI pattern just doesn't work for browsers anymore.
I think that you open more than one tab right away makes you bit of an outlier. I don't know if they collect this kind of data, but I would be very curious about the median number of tabs is among their users. I suspect most people use their browser in a very serial way. Check Facebook, check email, read the news then walk away.
If that's true, then maybe it's just a matter of priorities and constrained resources. Maybe an extension is where changes to the basic tab functionality belongs.
You've clearly put a lot of thought into this. Have you ever thought about building an extension?
If I'm going to build something I'd be building a new browser. What I'm proposing fundamentally breaks the architecture of modern browsers. To achieve what I want, the first thing I have to do is remove any visible notion of tabs. There is exactly 0 browsers on the market that expose an extension for me to hide the tabs directly.
No one can look at more than one thing at a time, nobody needs tabs, what they need is a browser that fits their workflows, and a browser that exposes an API such that third parties can experiment with page groupings and orderings.
Tabs are for workflow; they're a completely different modality for working with a browser, and it's not about MRU.
The reasons we open multiple tabs is primarily for cross-referencing, and secondarily "preview"
You've completely made this up and it bears no relationship with how I use browser tabs. Many of my tabs are open for months at a time, often whole trees of dormant tasks, collapsed.
As to screen real estate, I haven't maximized my browser window since about 2005, when I moved to a wide screen. Almost all pages don't render well to a wide screen format since it's hostile to reading, and for those that do render well, typically videos and presentations, full screen is better.
> When I have about a dozen tabs opened on Chrome, all I can see is a tiny favicon.
Which is precisely when tree-style horizontal tabs are better. Because you can see the entire title of each window, their nesting, and easily scroll through the list.
Why exactly would I want to have a side bar that has rows mirroring the exact same columns on the tab bar? To band-aid this problem, all Chrome has to do is implement tab preview on mouse hover like Opera has done years ago and now Firefox.
The Tab Groups functionality (later moved to an addon) and currently not supported in FF Quantum was better than just tabs. Made things much more manageable.
'Ctrl + Shift + E' and you would get a thumbnail view of all open tabs. Worked great while it lasted.
> I don't really need an actual side bar that takes up screen space
The good thing about side bars is that they don't take up screen space. Websites usually have large amount of horizontal space left for you to use.
The good news is that it is really easy to organize yourself when you use something like Tree Style Tab. You're obviously burnt out using horizontal tab in chrome and believe the situation is hopeless. I suggest you start using Tree Style Tabs for a few days or a week and realize how amazing it is.
It is really for a reason that it is the most beloved plugin for any browsers invented.
And you may find yourself
Behind the wheel of a large Microsoft Mouse
And you may find yourself on a beautiful PC
With a beautiful Window
And you may ask yourself, well
How did I get here?
Note that the shortcut is equivalent to hitting the "Tree Style Tab" button in the main toolbar, so if you're like me and want to show the tabs all the time you don't need to worry about having a keyboard shortcut.
If you do need the shortcut, AFAIK your only option is to edit the extension. Modify the manifest.json file in the root of the XPI around:
Unfortunately, actually installing your changes requires a decent amount of arcane incantation (particularly if you're not using Nightly) -- https://developer.mozilla.org/en-US/Add-ons/WebExtensions/Pu... describes how to do it if you want to go that route.
The one feature I've missed in the new version is click-drag across the close buttons to close multiple tabs in one fell swoop. Tbh I'm not sure if it was TST or multi-tab handler extension that used to provide that feature.
For the record, I don't like the chromeification of Firefox at all. But, if they force me to use Chrome anyway, I might just as well use the original. I expect them to be bought by Google any day now.
The last big features missing from Firefox are tab tiles and the ability to "run" an executable download directly. The WebExtensions API does not allow either of them anymore.
