I can tell you exactly what made me use tabs instead of bookmaking sites: opening a bookmark means making a new connection, new round-trip, loading, parsing and rendering a full page. Having a page open in a tab means it's already there, in exactly the state you left it in. It will still be there if the Internet goes down. It will still be there after it comes back up. It will be there tomorrow, it will be there next week, after hibernating and waking up your computer a dozen times in between. It will be there even if the page is already removed on the server.
The difference between a tab and a bookmark is like between a book on your desk vs. in a public library. In the latter case, when you need it, you have to go there and fetch it - not only taking time, but also possibly failing, as it might not be there anymore.
Is it a perfect interface? Not really. What I'd really like would be this, + in addition a bookmark feature that actually freezes and snapshots the page I'm on - to get the same behaviour as tabs, only persistent. It shouldn't be hard for browsers to do, given that IME at least some browsers displayed this behaviour at some point when restoring pages after browser crash.
I think this really hits the nail of why I am a tab-collector. If we could store “state” as part of the bookmark instead of just a URL that would go a long way.
But additionally, discoverability of bookmarks is also lacking, flipping through tabs is fast and easy but in order to do that with bookmarks you have to know what the domain and title of a page is/was. If you could “preview” a bookmark then it might help a fair bit.
It baffles me that bookmarks and history browsing in current browsers is still so terrible. In some cases it feels like it has gotten worse compared to a decade ago. Browser history certainly has, because for some stupid reason Chrome deletes your old history. I have no idea why it does it, but it's certainly not to save space.
I think the bookmarks UI has gotten so bad (or stayed stuck where it was instead of improving)... because so many people use tabs for what they could be using bookmarks for. A really well designed bookmarks UI might be able to do even better for them, but with tabs being 'good enough' (and better, for their needs, than the bookmarks UI they had historically when they started using tabs)... attention to bookmarks UI languishes.
A good bookmarks UI is a hard problem, and there's no guarantee you'll succeed at getting people to use it instead of tabs, so the cost-benefit-risk isn't motivating.
If someone did want to create a 'next gen' bookmarks UI... I'd start by intensively studying how and why people use tabs. :)
Bookmarks used to launch in a little docked file browser style window. It was ugly, but it was easy and quick to organize. You can still get it to behave like that, but it takes a lot more effort so, it has definitely gotten worse in my opinion. Certainly it takes more effort to organize.
Really, what they should have is a "Snapshot" option and a bookmark option where the snapshot saves a static local copy of a site and a bookmark saves a link. Bookmarks should have RSS integration so you can optionally sort by recently updated. History needs a meaningful search option and a way to group/nest by domain. We haven't even tried to pick the low hanging fruit.
Bookmarks are however, antithetical to Google's worldview and business model which I think drives some of their languish given that so much of the internet now relies on Chrome.
> for some stupid reason Chrome deletes your old history. I have no idea why it does it, but it's certainly not to save space.
The pessimist in me would say that's to make you go back to Google and search for it again in the hopes that they can show you more ads on the results page.
That's the way I've always overcome the problem of getting to bookmarks also. You give up a little real estate to keep the bookmarks bar visible, but it is a good way to organize things by subject matter, project, urgency, etc.
That being said, I'd love a configurable cache that would allow me to open a group of bookmarks to exactly where I left off on each page (or where I had frozen the data to keep a static record.)
I don't understand that Firefox shows me a great start page with my recent visited pages with thumbnails but my bookmarks are still an obscure text list.
I've searched for an extension that would turn the bookmark sideboard into a full page with thumbnails but I haven't found one. (And I am too lazy to start such a project. I also assume that the Firefox team will light up a turn the bookmarks sideboard into something like the start page.)
Adding an option to store a copy of the page in the bookmark would make this a great usability enhancement.
booksmarks as implemented lock down a particular url, whereas tabs keep the state of what i am working on (reading)
a tab may contain the the last article i read on a blog (for example it keeps the last day when i read hackernews: https://news.ycombinator.com/front?day=2019-06-19
if i am busy for a few days, then i know exactly where i need to continue reading to catch up.
works for web comics and anything else posted in serial form.
i DON'T want to bookmark these pages. because every time i read them, the url changes. i have found some extensions that update bookmarks as i go along, and that worked well to track webcomics, but ultimately it always came down to keeping a set of tabs open, because that is really the only way to track your progress, including a history that allows me to go back.
another use is actual research. i may research a problem, that takes me to stack overflow, then i follow a link to a referenced article, etc...
the tab keeps track of where i was when i last worked on the problem, and allows me to continue effortlessly.
if i bookmarked these, i'd have to remember to remove the bookmarks after i am done. with tabs i can just close them.
That's an extremely difficult problem for browser makers to solve in a generic way: the state of client-side scripts might not lend itself to such a snapshot, and the browser has no way to account for changes that have been made server-side in the meantime (including new versions of the page content, scripts, stylesheets, API endpoints, etc.). Your proposal reminds me of Java object serialization, which I've rarely seen used in an effective way (and only when the code and runtime environment are tightly controlled).
Webpage makers already have the tools they need to do this: the page URI can be manipulated client-side to store state. And they can specify which content should be cached for use in the future without a trip to the server—including full client-side app functionality.
Server side changes are moot (1), because in order to properly do state saving, you have to save more than the assets in a simple cache, but also the entirety of the DOM as of that particular moment and the state of the js engine.
I couldn't even imagine supporting that in an app I wrote, much less in a way that works for every website everywhere!!
(1) moot in the sense that it ought to act to server side changes much like any static pages would become out of date when backend updates are made. If the api endpoint goes away, then it errors, it's not the responsibility of the browser to keep the app evergreen, imho
> because in order to properly do state saving, you have to save more than the assets in a simple cache, but also the entirety of the DOM as of that particular moment and the state of the js engine.
Sure, it should do precisely that, IMO. Serialize it and dump it. All of this is in browser cache AFAIK anyway, so it's not like it's some completely novel programming exercise.
As for keeping up with the server side - I don't care. The "sync to the server state" button exists, and is called "Refresh" button. Nothing needs to be implemented by website author. In fact, I'd prefer if websites didn't have a say in this at all - or else someone will find a reason to screw with the mechanism.
If the server side has changed then I can simply refresh the page if I want the newest page. I'd like the page to be exactly what it was when I closed the browser. I've left computers running for days just to save the state of a single browser tab. I suppose a better way of doing it is to only use browsers inside a VM and them just save the state of the VM.
About websites: I don't want to have to rely on website makers, because they invariably won't do it and it won't work.
That only works if the page is relatively static. Add in any functionality that requires a round-trip to a server, and things get immensely more complex.
You can press the Refresh button and suddenly things fix themselves. The utility is in having a snapshot of the site looking exactly as when you left it, with all the benefits of it not being rendered to a image, and with a high probability that at least some interactions still work.
I would be happy for a render to PDF that had some basic parameters around when to perform a render, max storage to be used, aging, or if not that, first in, last to be deleted as new comes in functionality.
Think of it as a cache of things seen. Bonus points for hyperlinked media assets. Double bonus points for those links being to cached media assets.
Somehow these are all reasons why I don't like tabs.
Firefox has pinned tab feature and I liked the idea. I could get rid of my go to bookmarks in the bookmark-bar and save a bit space.
But tabs behaved SO differently from bookmarks, I just couldn't stand it. Everything felt wrong and outdated and I had to constanly refresh stuff after I switched back to a tab.
Safari actually has the ability built-in to export a page as a PDF. Which you can then open in browser. It can be damn useful sometimes because it preserves all of the formatting and everything that's awful to try and preserve when you save the web page itself.
Say that tabs are not the right metaphor is like saying that spreadsheets are not the right metaphor. There are better metaphore for every single case, but it's a very good one for most cases.
Why would I, when I have tabs, and know that if I force-kill Firefox or Chrome, I'll get the "Session restore" screen the next time I run the browser? That's essentially "local bookmarks", but with much better UX.
(Yes, I do really force-kill browsers without mercy when rebooting. It's an effective strategy for ensuring I'll get at least some of my state back, and is more convenient than "session manager" features that keep coming and going.)
I'd love to see more publicly available research and experiments in changing the desktop metaphor. We see stuff that makes tweaks around the edges (new context menus, unifying file/cloud presentation, tiling window managers) but very little that radically changes the way we actually interact with computers.
For my ideal work desktop, I want to have tools assigned per task that I can easily assemble until that task is achieved. As a spitball idea: Cmd-space, type in "research open sourcing my kindle", and get given a browser and some sort of note collector. Collect the various things from the session into my note, refile the entire stack as a project, and move on to the next task. In the future, I can access the actual project and receive a workspace with terminals, an IDE or whatever else I need to Get the Thing Done, archived versions of the sites I opened, etc. Ideally, the environment would be designed to reduce distractability.
That still doesn't feel like a complete idea, but every use case the author has for tabs -- "this is a quick session", "I don't want to lose this page", "I need to remember this in a week" -- seem like they'd be very hard for browsers to do well. Instead, a browser more closely integrated with the OS (which might go a way towards fixing Electron resource consumption to boot) and a uniquely designed desktop environment might be better.
While it is only a design concept, Mercury OS [1] might interest you. When I first came across it, I was blown away by the idea. It basically makes the user intent first class rather than the application/program. What I mean is that instead of finding a program or application to execute your intention indirectly, you declare your intent to the OS. Then the OS maps your intent to a series of actions which are executed by hooked-in pseudo-functions. It erases the boundary between applications by being almost like an interactive visual version of a declarative programming language. You should read the article rather than trouble yourself with my explanation though...
The workspace concept (Eclipse, VSCode, Visual Studio solution etc.) works well, but is often hated because it is at odds with the OS file system hierarchy.
Making workspaces a first class OS citizen would be great, but the problem is with mutually referential hierarchical projects, having to include shared network folders in a corporate workflow etc., which break the metaphor again. Maybe an n-ary taggable workspace association flag on FS entries would work (Reiser FS?).
what do you mean by "at odds with the OS file system hierarchy"?
the file organization in a VS project can be totally abstracted from the underlying file structure, or it can be a 1-1 mapping. if I'm using a supported version control system, I might never need to care where the files live on disk. on the flip side, the fact that every file has a physical location mean I can also use whatever unsupported version control I want.
the OS filesystem is a pretty stable and relatively simple interface. I'm not sure how it could be made more intuitive without sacrificing flexibility.
Personally I am a big fan of haveing file system hierarchy dictate workspace heirarchy. I understand that there are cases where it might be nice to have a separate structure defined for an IDE, but that often comes with the cost of not being able to seamlessly use filesystem operations to edit the project directly. It feels more *nix-y to use the filesystem for this.
I thought it one of the really interesting parts of the Sets experiments that the Windows team was working on (and ultimately seem to have cancelled). They fell straight into this rabbit hole of "sessions" and defining what they mean to a user as soon as they tried adding (heterogeneous) Tabs as a native UI element to every window title bar. A lot of the dev talk on what to expect with Sets was on trying to surface user-meaningful activities to the Microsoft Graph and thus to the Windows Timeline thus making those groups of tabs meaningfully relevant to a user ("these are the tabs I used to file my expense report").
It is definitely a hard problem to solve, especially if you are hoping for software to do a lot of that aggregation automagically "reading people's minds". It was really cool seeing some very smart people on the Windows team trying to build it with Sets, and maybe all the more disappointing that the idea didn't make it "out of the lab" for its grand attempts at trying to build a more complicated/nuanced/powerful model of how people interact with groups applications at a time.
I like the idea that the job of the OS (or really, the DE) is to respond to why we're using our computers. Each application should have its own job, but the DE's job is to help us achieve our intentions.
For my ideal work desktop, I want to have tools assigned per task that I can easily assemble until that task is achieved. As a spitball idea: Cmd-space, type in "research open sourcing my kindle", and get given a browser and some sort of note collector. Collect the various things from the session into my note, refile the entire stack as a project, and move on to the next task. In the future, I can access the actual project and receive a workspace with terminals, an IDE or whatever else I need to Get the Thing Done, archived versions of the sites I opened, etc. Ideally, the environment would be designed to reduce distractability.
I'd willing to donate money, work, or both. Is anyone working on something like that?
Have you tried https://workona.com/? Is a chrome plugin so limited to web tools and resources, but for me and the use case you describe, it’s a total game changer.
That's exactly the point of using HyperLook as a window manager, which we were discussing at Sun in 1991 before they canceled NeWS.
HyperLook, which was inspired by HyperCard, but written in PostScript for the NeWS window system, let you build specialized task-oriented user interface by assembling and customizing and scripting together components and applets into their own "stacks".
>HyperLook was like HyperCard for NeWS, with PostScript graphics and scripting plus networking. Here are three unique and wacky examples that plug together to show what HyperNeWS was all about, and where we could go in the future!
>HyperLook (which ran on NeWS, and was like a networked version of Hypercard based on PostScript) was kind of like looking outwards through windows, in the way Rob Pike described the Blit (although it was architecturally quite different).
>I think his point was that window frames should be like dynamically typed polymorphic collections possibly containing multiple differently typed components that can change over time (like JavaScript arrays), not statically typed single-element containers that are permanently bound to the same component type which can't ever change (like C variables).
>With X11, the client comes first, and the window manager simply slaps a generic window frame around it, subject to some pre-defined client-specified customizations like ICCCM properties to control the window dressing, but not much more, and not under the control of the user.
>With HyperLook, the "stack/background/card" or window frame came first, and you (the user at runtime, not just the developer at design time) could compose any other components and applications together in your own stacks, by copying and pasting them out of other stacks or warehouses of pre-configured components. [...]
Here are some ideas about HyperLook and other systems, that could be applied to Sugar:
>[...] My ultimate fantasy gui environment would support plugging in different script editors, with not only easy-to-use but also advanced visual programming interfaces, tailored to specific tasks and skill levels (like handling and sending events, solving constraints, processing signals, playing music, controlling robots, blinking lights, dancing bears, etc). [...]
The trouble with bookmarks is they take conscious effort to create and maintain.
The trouble with history is that it's a soup of every random page you've ever visited.
The trouble with tabs is they're a collection of web pages you at one time thought you might get back to, roughly sorted by the time in which you first opened them.
The way history is always implemented is also not helpful.
Say you browse sites A B C D E, then later you go back and browse D, then B.
History seems to be recorded as A C E D B
But recording A B C D E D B might be more cluttered, but more helpful as well. It shows you the decisions you made to browse from one thing to another.
This compaction method also seems to happen across days, destroying the list of everything you browsed on tuesday in order.
> Say you browse sites A B C D E, then later you go back and browse D, then B.
> But recording A B C D E D B might be more cluttered, but more helpful as well. It shows you the decisions you made to browse from one thing to another.
That's personally why I have been using Tree Style Tab (for about a decade now)
Firefox's implementation of this annoys me. If I'm trying to find a page I was on before, I sort history by date (obviously) and go back to around the time I was on it to start clicking.
...which drags the website I clicked on to the top of my history, changing its temporal position and making me have to find the time period I was looking at before, which now looks textually different.
There's no reason why there couldn't be multiple views of your history, if it's recorded with fidelity. The list of page views by time is often too noisy depending on what you are looking for, and the compacted view might be more useful. As you rightly point out, sometimes that page view log is what you want.
Really, the biggest use-case for browser history in its current form, seems to be to get people in trouble when nosy family members go peeking at it.
Completely agree. I don’t really understand the motivation for this behavior—surely they are not trying conserve disk space? I’d really like the history to just be like a log file. Even better if I could search within a specified time window.
It sounds like a close-to-correct metaphor is a list of tabs associated with temporary bookmarks per tab (sorted in order first opened in the tab). That is, when you visit a new page, by default your current page does not get added to the list of bookmarks, but from the right-click menu/middle click you can choose to have your current page added to the list.
>2019-06-14 Application status is Expired - Lifetime
Wow, thank you for catching that and pointing it out, so I can celebrate just in time! How did you happen to coincidentally run across that fact today? ;)
(I have always believed was illegitimate and invalidated by a lot of prior art, but I didn't realize that its death date was imminent.)
Have all tabbed applications and browsers been paying royalties to Adobe? If not, have they technically been in violation of the patent this whole time? What were the real-world implications of this patent?
>"Adobe filed a lawsuit against Macromedia® alleging infringement of our intellectual property, specifically our patented tabbed palette, which is a key user interface element and method invented by Adobe and incorporated into our products."
>"Adobe's tabbed palette patent is Adobe's method of displaying and working with multiple sets of information in the same area of the computer screen. The patented invention allows users to customize how the functions in the product are organized on the workspace. This invention was a significant leap forward for customers' productivity and personalization of the interface."
In 1989 the NeWS PSIBER Space Deck supported tabbed nested windows that users could drag around separately on the screen or group together on a visual representation of the PostScript stack, use "direct stack manipulation" to push and pop and rearrange the PostScript stack by dragging, and use pie menus to switch the side of the tab, pull-out to adjust layout parameters, issue commands, etc.
Ironic that it was part of a visual interface to Adobe's own PostScript programming language (in Sun's NeWS window system implemented by James Gosling), so they had to pretend not to know about it, in order to fraudulently obtain their tabbed window patent through willful blindness.
I know for certain that at least one high level person at Adobe was aware of it, since I corresponded with Glenn Reid at Adobe about PostScript quite often before during and after its development, and credited his PostScript book and his work on Distillery in the references of that paper.
>Adobe Systems; Reid, Glenn C.: PostScript Language Program Design (The Green Book); 1988; Addison-Wesley Publishing Company, Inc., Reading, Mass.
>Reid, Glenn: The Distillery (program); available from ps-file-server@adobe.com
There's not much incentive for a company to pay Adobe licensing fees for such an obviously invalid, fraudulent patent. But it costs a huge amount of money and time to challenge a bad patent like that and clean up the hot mess of bullshit that Adobe dumped on the world of user interface design, far beyond my means or the means of most non-giga-corporations.
I used to have many tabs open myself. Not so much for different tasks, but more for "i don't have time for that now, I will read it later".
Now I use a firefox extension called "Bookmark Stack" for that. I can easily add the current page to a stack and later restore it. Since it is a stack, the most recent entries are always displayed first. Since these are bookmarks, syncinc is also easy.
I wrote that extensions myself, but there are others like it. Since using it, I rarely have more than 10 open tabs. That may not suite every workflow, but it definitely helped me.
Yes, the stack is growing (currently somewhere around 200). However, I think in terms of resources it's better to have a lot of bookmarks than to have a lot of open tabs. Also there is almost some kind of gamification to try to reduce the number.
There's an old joke about four-wheel drive on cars: instead of getting stuck right when you go off road, it ensures that you'll get pretty far into the woods before getting stuck. IHMO that metaphor applies here. Like the Web, Google, and even the WIMP interface before them, tabs were invented to manage information overload but "succeeded" only in pushing the problem out a layer:
"The most sophisticated and efficient users of Google today find that “trying to navigate the growing deluge of data [just within their scientific field] has become a second job.” The United States’ library of scientific papers for health and medicine adds new ones at a rate of two a minute, year round. Like Memex and the Web, Google has solved the problem of published work burying itself unread by pushing it out to a vastly wider layer, one we will perhaps never manage to tame. We have not escaped [Vannevar] Bush’s quandary:
Those who conscientiously attempt to keep abreast of current thought [find that] truly significant attainments become lost in the mass of the inconsequential. … publication has been extended far beyond our present ability to make real use of the record."
The history here, is that Opera introduced tabs, and they were cheaper in terms of resources. Tabs in Windows were cheaper than many windows. Then tabs became a sort of replacement for Bookmarks, because bookmark UIs are so so bad. That's my take on it.
Gnome 3, started out suggesting that it would be a semantic desktop, allowing groupings associations of activities. That fell by the wayside, so the best thing left is to use something like virtual desktops to separate projects/workflows. But then you are pushed into multiple windows likely with multiple tabs.
I use multiple browsers on multiple devices, and bookmarking isn't that nice. I see why people end up using browser syncing and end up in vendor traps. On desktop, I can easily copy paste urls. Perhaps into my own mini-shared db for read it later style groupings. On mobile copying URLS is hell for me.
Imagine throwing tabs away. If windows were cheap again how could you best use a Window manager to manage them? It feels like this should be at the heart of your desktop. It would need to be more than 'recently used/accessed'.
I have tried running Firefox with the bookmark manager open and the history, but it has been really poor performance wise, and it's just not a nice interface.
Looking at my tabs today, I have a small grouping of shopping tabs, of things I might be interested in ordering. From about three weeks back. I know if I throw these into bookmarks, and even tag them, they'll still be lost. I should probably have dealt with that there and then. I tend to build up tabs in Firefox, once I hit the width of the browser with tabs and they scroll it becomes UI hell for me. So I try to confine to about 12, but that's easier said than done. And as said I run multiple browsers, at least four on mobile and four on desktop. Throw tabs into the mix and it gets ugly. But if this was reduced to something like a simple list of URLs, with some grouping shared between machines that I could relaunch easily, probably in the same browser I was using previously they'd feel much more manageable. And less crufty.
Going back to the shopping example, that's task focused. And it might be accompanied by a few notes written in another app, and I might want to share that with my partner via email etc. Associations have to beyond one piece of software.
> Imagine throwing tabs away. If windows were cheap again how could you best use a Window manager to manage them? It feels like this should be at the heart of your desktop.
Sounds like you are describing surf[0] and tabbed[1] from suckless.
This seems like a good place to ask. There was a Firefox extension called (edit: not sure if it was Tab Mix Plus or something else) that was discontinued after Firefox 57. It was super productive and still haven't haven't found anything like it to replace it.
To be clear, I'm talking about grouping tabs but I'm not talking about the current Firefox feature called "Tab Groups". You had a row of tabs that showed the web pages open in the current group. And you had a second row of tabs that allowed you to switch between tab groups. You could close a tab group and close all the tabs within the group.
Has anybody found an alternative even if it's in a different browser? I dislike having multiple browser windows open.
Tree style tabs (Firefox) works in much the same way, assuming you have a parent tab for each group.
I routinely start a research task from one tab, say a search query, open all interesting results in sub tabs and close the parent tab including all children (ie the group) once I'm done (or I collapse the group when I intend to resume later).
Browsers are moving towards more restricted extensions and don't allow the unlimited UI edits that Firefox used to allow so it's unlikely that an extension would allow that exact functionality. Maybe one of the Chrome/Firefox forks could add it as a base feature.
It does bring back memories, I wrote what I think was the first extension that did that type of tab grouping. Was something like 12 years ago and even had the name "Tab Groups."
Man, okay so clearly we all agree there's a problem and it sounds like we all have different ideas for a solution. One of my pet ideas - that I'm yet to see a real implementation of - is a Hacker News comments section style history.
Each new session would start a new comments section.
Each new manually opened tab would be a top level 'comment' (with timestamp).
Each pageload would be a paragraph in that 'comment'.
Each new tab opened via a link would be a 'reply' to that comment quoting the 'paragraph' it was opened from.
That seems like a good metaphor to me because it solves the traceability of "How did I get to this page". It doesn't solve the UI of tabs though, though maybe it makes it less necessary to keep tabs open.
I use it, but not often enough to make a real dent, my issue is that it isn't integrated into the browser, it's a feature that with some UI tweaks could really make everyone a poweruser, the concept itself is powerful but the un-extensibility of chrome (and browsers in general) holds it back from being a true first party app.
Agreed. And it's one of the reasons why I dislike the trend of the browsers being increasingly locked down and non-modifiable. Soon, even to recreate this plugin you'll have to run a parallel fork of a browser.
I think it's equivalent to having books open to a page and set to the side of a work area.
Whereas a browser bookmark is like it's analog cousin, a mark stuck in the book and returned to the shelf.
Tabs allow for better singular focus than stacked windows. I use desktop workspaces/multiple desktops in much the same way I use tabs. If I have multiple monitors, I'll put a group of related applications together in a workspace, but rarely will I put more than a single application on a monitor.
Eg editor on one screen, terminal on the next and a browser (UI work) or DB interface on the 3rd. Which is the same way I organize my kitchen when cooking. Recipe on one side, prep in the middle, output on the other.
Had struggled with this for 8+ years with usually 200+ tabs open, even started coding my own extension for this. And tried at least 30 extensions.
Finally settled on https://workona.com/ . It's not exactly perfect and I disable the tab cache feature via settings to basically open each workplace in a different window instead of replacing an existing one. But otherwise, it's perfect.
"I usually have at least 6 browser windows open with at least 10 or more tabs each. I wish there was a better way to do it."
Back when Opera was an innovative browser, it had a really useful tab grouping feature which let you achieve the same thing as you describe but without requiring a separate browser instance.
you mean the different "spaces" in the video for work, thesis, etc.? i don't think that's AI driven, it looks like the user manually puts a tab into a space and tabs open in the current space by default.
That's pretty much what I initially thought/hoped Firefox 'containers' were. But it's not, with those you can have X open tabs in Y different containers all in the same window.
Wow, this is a cool concept. Efficient tab management especially is a feature that should be baked into browsers, there should be no need to install an extension for that.
I used to use an extension called Tab Wrangler [1], which would close tabs after being inactive for a period of time. I felt this dealt with the problem of too many tabs surprisingly well—Tabs would quietly disappear as I forgot about them. If I needed a tab back (which was more rare than I would think), Tab Wrangler could show me a list of tabs it had recently closed.
I've switched to Firefox a while back and unfortunately cannot find an equivalent, so I'm back to manually wrangling.
I was hoping this would happen when tablets were introduced. But apparently, for some reason, the idea of infinite, zoomable canvas for interactive objects and annotable documents is alien to most software developers. All the products I've ever seen either a) allow viewing and/or annotating single documents, and/or b) use a very limited canvas with a zoom limit, and/or c) don't allow to put documents in it in other ways than rendered images.
http://web.eecs.utk.edu/~azh/pubs/Henley2018bDissertation.pd...
FWIW, for those interested, there's some discussion of the "Code Bubbles" environment which allows arranging working sets on a canvas. (IDK if this falls under "limited canvas" or whatever).
I think a big part of why this hasn't caught on is that the benefits of a layout which maps to a mental model are _generally_ less than the cost of arranging things.
I came to share basically the same idea. I know there are extensions but I want some kind of workspace/session support built in so the browser can do stuff like cache tabs in a workspace so I can load a workspace and resume instantly. Switching workspaces should be quick with maybe a UI like ctrl p in sublime text or VsCode or even like the gnome dash. I would pay for a license to use a browser that focuses on being a productivity tool with features like this.
Been really enjoying workona as well. Have found myself recommending to folks on my team. I organize my tabs into their “workspaces” which map to the strategic and tactical projects I’m juggling. Helps me switch btwn contexts
I actually changed to Chrome only for this. I have been looking for months for something like Workona for Firefox but to no avail.
I used Firefox for years and I want to continue using it, but for my way of work (multiple tabs open in several spaces and changing computers often) Workona is just great, nothing comes close.
Except there's no way to sync the state back. Say you open another tab in that window. If you bookmark that window again, you end up with dupes. There's no coupling between a window and the bookmarks folder that spawned it.
I recently found in surprisingly difficult to explain the concept of browser tabs to my relatively computer unsavvy 72 year old mother. She can do web searches, read online news articles, and watches Youtubes, but browser tabs were a struggle for some reason.
I try to teach my mother how GUIs work and it's really hard. It's hard enough to teach her that pressing "X" closes the window, but then she opens something like a web browser and it has "X" on every tab, which gets her confused. And worst of all, she can have email client on top of the web browser, which adds additional "X" to think about.
Perhaps keep them clear of the idea. And stick to Windows. Assuming they can do Window management...
I've an 84 year old family member, who still doesn't get tabs. Heck they don't really get the Web. And are kind of stuck in their Web Email client. I got an alarming call one day saying they'd hosed all their emails. But it turned out they'd failed to log into their email account, ended up creating a new one, and then wondered why their inbox was empty.
...or good old-fashioned recipe cards. Most card files were arranged with tabs, but a box of recipe cards would be more than familiar to anyone my age or older. (You'd pretty much have to have lived out of cans and/or on restaurant meals to have avoided them.)
I can’t speak for anyone else, but I find that my usage of the web is more like a mind map. In other words, a tree or a directed graph.
I open a tab to accomplish something. From there, I reach out in some primordial/human-brain breadth first search of links in order to accomplish my goal, branching off of branches and coming back up to some parent only to branch again.
For people like me, a linear set of tabs is indeed the wrong abstraction for the way I use the web.
Here's "iLoci", an iPhone app I developed in 2008 that integrated the web browser component and url bookmarks into a mind map or "memory palace" based on the "Method of Loci", that lets you arrange links, pictures, text and other interactive content into an easily editable 2D navigable maps (the gestural equivalent of a network of pie menus):
People use tabs as bookmarks and as history and back button. Users want bookmarks and history and the back button, but apparently the tab feature is better at this than the actual features.
Maybe browsers should embrace this instead of fighting it. Make tabs better suited as history and bookmarks. Users reflexively middle click and then go to that tab because they don't want to lose the page. The problem is that you lose the pages that you visited in the first place. Maybe the default behaviour of a link shouldn't be to navigate away from the current page and hide the current page behind the back button (where you'll eventually struggle to get back to it) or in history (which also isn't easy to get back to).
Maybe the default behaviour of clicking a link should be to open a new tab, so to speak, but to focus it while attaching it as a child to the tab it came from, thus creating a tree of history. Closing a tab just means hiding that part of history. Viewing history is just unhiding those closed tabs. Bookmarking is just reorganising that tree of tabs.
Trying to solve a problem that isn’t actually a problem. Tabs work fine. “But your using the tabs sort of like bookmarks” is like complaining someone is using a fork/knife combo to eat pasta and wanting to therefore redesign a knife that works more like a spoon.
The article’s premise is mostly an arbitrary value judgement that a high number of tabs is negative in some way (“ridiculous number”). Notice that there is no concrete definition of what the drawback would be. There’s a strong element of rule-seeking and cleaning instinct at play. Not that there’s anything wrong with that ;)
Yeah, I never understood the pasta spoon thing either. I just don't see how it helps anything when you can simply use the plate as a backstop for the fork to roll the pasta correctly and not have hanging strands.
Apparently in Italy, the spoon rolling technique is only used by children.
>I don’t think I would remove the tab altogether, but just show me the last few things, treat them as emphemeral and provide an option to explore my current and previous session(s).
I would never use a browser that did this. I have many tabs open. I've even started using several (!) browser windows and profiles to have more. Being able to easily go back to previous stuff I've opened is really nice, even if it takes me a year. If I had to use bookmarks instead, then they'd probably stay there. I have many bookmarks that are about 10-15 years old. I don't think I'm ever getting back to those, because they aren't a constant reminder like browser tabs are.
Tabs on the left and right edges of the windows work much better than top or bottom tabs, because text is wider than it is tall, so you can fit a lot more tabs, and actually read their labels.
Users should be able to decide which edge and position each tab is attached to, and move them around and group them as desired.
These are the tabbed windows I developed for the multi-window version of UniPress Emacs 2.20 in 1988:
"As if all that wasn’t enough, there’s also the matter of tabs. Tabs are a couple of decades old now, and, like much of the rest of the desktop and web environment, they were initially thought up in an age where the predominant computer displays were close to square with a 4:3 aspect ratio."
Tabs are at least three decades old now, and they weren't originally restricted to just one edge of the window (2018):
> Tabs on the left and right edges of the windows work much better, because text is wider than it is tall, so you can fit a lot more tabs, and actually read their labels.
Yes, I find that more than a handful of tabs at the top of the page are hard to use.
The wonderful "Tree Style Tab" add-on for Firefox puts your tabs in a tree at the side of the page:
The hierarchy represents parent-child relationships between pages: if you open a page from another page, it becomes that page's child in the tree. You can drag tabs around to rearrange them, collapse/expand/close subtrees, etc.
It's one of the very few add-ons I can't live without (the others being uBlock Origin and uMatrix).
Hi Don! Tabs aside, NeWS got so many things right, and not just at the UI but the underlying idea of code instead of bitmaps.
I wonder if we were going to start afresh with Wayland/X11 now, bare hardware, what place do you think the NeWS lessons would have? Would a modern NeWS be on top of PS or is there something better now?
Simply use a standard JavaScript/WebAssembly/WebGL/Canvas/HTML based web browser as the window system itself! And use WebSocket/SocketIO/RTP/HTTP instead of the X-Windows or VNC protocols.
Microsoft kinda-sorta did a half-assed inside-out version of that with Active Desktop, but Internet Explorer wasn't powerful or stable enough to do it right, and it didn't eliminate and replace the whole Win32 / MFC layer, which misses the main point.
There was recently this discussion about a browser based Mac window manager project (now offline), and there have been others like it (Slate), but the ideal goal is to completely eliminate the underlying window system and just use pure open web technologies directly on the metal:
Show HN: Autumn – A macOS window manager for (Type|Java)Script hackers (sephware.com):
>I was also quite inspired by Slate. Unfortunately there hasn't been any activity with it for about 5 years or so. It's great you're picking up the mantel and running with it, because the essential idea is great!
Opera had side tabs (as far as I remember you could freely choose one edge to display them on). I really miss the old Opera browsing days, it had great functionality such as more keyboard-oriented browsing. E.g. you could bind a key to "fast forward" through pages of a site - it parsed out the paging links to simulate clicks, worked in most forums back in the day.
I believe Vivaldi has both of these features. Unfortunately, it seems a little bit more resource intensive than Chrome or Firefox on admittedly low-mid tier hardware.
What I would love is an extension (either FF or Chrome) that allows me to reuse an open, matching tab when I choose a bookmark. For example, I find myself opening a bunch top-level Google News tabs. In a window with 40 tabs, at least 6 or 7 are the top-level Google News page (or NYT homepage). I would love to be able to select my bookmark and have it find an already existing tab that contains an exact match and bring that tab forward.
This would reduce (for me) a large amount of tab flotsam. Maybe there is already a way to do this?
Using Keyboard Maestro on a Mac, I have keyboard-based shortcuts for opening specific web pages. If I invoke one of those shortcuts for an already-open site, it is brought to the front instead of being opened again in a new tab.
If you use a tiling WM, the immediate question which comes up is : can I switch this to a clean single-desk view, or is this neccessarily a tiled plane of n things, and this is one of n?
Tabs, are an attempt not to have to switch desktop, or tile the plane. Its a "stack" metaphor, with the least cost exposure of the underlying stack, as a shuffled hand of cards fanned out.
Hand of cards, vs one piece of paper per desktop but many desktops, vs I cover the one desk, but all things have to balance the space.
Automatic bookmarks based on the signals you generate from browsing, history laid out as a tree, tabs remembered with temporal grouping so you don't have to keep them around. Notes and tags, with everything searchable, shareable, and easy to publish.
I'm not clear why the author thinks that his dislike of others usage patterns means that there's a problem in need of solving.
If his friends and family are happy with using tabs that way, this seems like a very poor reason for browser vendors to change things up to show them the light of his workflow.
Browser tabs are the wrong metaphor when used with only 1 window. If you keep unrelated tabs in different windows it makes way more sense and the number of tabs stays relatively low.
An obvious use case is to have a window for your work and another one for your personal stuff. You can also open a new window when you start googling stuff about a specific topic (let’s stay “how to do x in Python”). Then, when you got your answer, you just have to close the window with all its tabs.
I'd think part of the challenge of coming up with the UX of "the right metaphor" is the discoverability and intuitiveness of the solution.
Tabs are really obvious. You can do a bunch of stuff with tabs and there UI is so apparent that it doesn't get explained.
Apparently the usability threshold for people with 100s of tabs open is (a modern browser's) bookmarking. Any superior solution would need to be more obvious than bookmarks and easier to use than bookmarks.
Huh, I always thought Opera was the first tabbed browser but according to wiki, BookLink Technologies' InternetWorks browser was the first, followed by Netcaptor three years later.
I daydream sometimes about having a browser where tabs behave more like buffers in vim or emacs - that is, with no tab bar or particular ordering on open pages, but with them just hidden in the background, with some kind of about:pages interface (like emacs ibuffer-mode) where I could sort/filter them and bring individual ones back to the foreground.
(Split panel view would be nice, while we're at it.)
There's a button on the top-left which collects the current tabs in the window and puts them aside. They can be restored later.
Personally, I've only used this by accident. But I could see it being useful for "I know I want to come back to these tabs later (e.g. tomorrow), but I want to close the browser / do other things now".
That sounds like a way to save a browser session that comprises of tabs. Perhaps a better UI. Many browsers have had the ability to group/tabs/bookmarks. But it's usually quite clunky.
>One thing that kind of interests me is that every now and then I spot someone with a ridiculous number of tabs open. [...] Maybe commenting on their poor organizational skills,
I often hover above 50+ tabs open. Every week or so, I catch up enough on the backlog to close some tabs and reduce it to less than 10. But I soon head down another rabbit hole of internet research and it's back up to 50+ open tabs again.
The bookmarks UI has extra friction because is a separate area that I don't want to "manage" and that effort isn't worth it unless the webpage is promoted in my mind to "reference status". Having dozens of open tabs is just an easier and a more seamless experience. I often cycle through open tabs with Ctrl+PgUp and Ctrl+PgDn similar to cycling through active windows with Alt-Tab. Bookmarks UI don't have a "cycle" mode.
Another problem with booksmarks is that the bookmark_links are not open. The booksmarks manager UI only show the title of the webpage but not what full page rendering is. Titles are often meaningless.
Of course, I could manually type in a better title when saving the bookmark... or... just simply not close the tab and consider it part of a "queue" to read later in a few hours or days. It's lazier and easier. Plus the full page already completely rendered visually reminds me why I wanted to read it later. Bookmarks UI obscure the _why_.
For the webpages I consider as reference material, I do bookmark those. E.g. weather page for a specific zipcode or a specific Youtube channel.
Examples of webpages being categorized into different mental organization buckets:
level 1: throwaway content: a trail of leaving a bunch of active tabs open is easier to get to than the friction of navigating the bookmarks UI
level 2: reference material that changes: bookmarks for things like weather page, current US Treasury yields, etc. Bookmarks are also useful for syncing/import/export across devices since bookmarking services sync at the level of "bookmarks" and not "open tabs".
level 3: reference material archived: webpage saved locally as .MHTML or .PNG -- to ensure I have a copy even if the url is lost from digital rot. Others might use Evernote or Zotero to archive webpages.
I do agree that Chrome squishing the tabs into tiny vertical slices makes it harder to use. I wish they had an option to show open tabs list in a dock on the side. This would work well for widescreen monitors since many webpages would show useless whitespace on the sides when the browser is full screen.
I don't maintain anything close to inbox zero, but on Monday when I come into work I'll close the ~50 tabs I have open and only leave 1-3 of what I'm literally working on that moment, first thing Monday morning.
#/media/File:HyperTIESAuthoring.jpg){kind=link}
I can tell you exactly what made me use tabs instead of bookmaking sites: opening a bookmark means making a new connection, new round-trip, loading, parsing and rendering a full page. Having a page open in a tab means it's already there, in exactly the state you left it in. It will still be there if the Internet goes down. It will still be there after it comes back up. It will be there tomorrow, it will be there next week, after hibernating and waking up your computer a dozen times in between. It will be there even if the page is already removed on the server.
The difference between a tab and a bookmark is like between a book on your desk vs. in a public library. In the latter case, when you need it, you have to go there and fetch it - not only taking time, but also possibly failing, as it might not be there anymore.
Is it a perfect interface? Not really. What I'd really like would be this, + in addition a bookmark feature that actually freezes and snapshots the page I'm on - to get the same behaviour as tabs, only persistent. It shouldn't be hard for browsers to do, given that IME at least some browsers displayed this behaviour at some point when restoring pages after browser crash.