Hacker News new | past | comments | ask | show | jobs | submit login
Refresh: concept for a new kind of web browser (refresh.study)
575 points by davidbarker 10 months ago | hide | past | web | favorite | 146 comments

This looks really nice, kudos! Here are my few (hopefully constructive) comments:

- The website looks great but quite heavy:

  29 requests, 17.66 MB transferred, Finish: 40.09 s

  Maybe see if you can re-encode `history-device.mp4` and `spaces-device.mp4` since they constitute the 78% of the traffic. =)
- I liked all the features you've introduced, but loved the way you've designed them!

- Kudos for getting the tagging right! Many developers expect users to tag their content, and tags were the next-big-thing at some point where there were tag-based file systems[0], photo and video galleries that expected you to tag each of your photos, and so on; but what really worked was automating the process of tagging! Everyone can agree that Google or Apple Photos are miles ahead of digiKam or Picasa when it comes to finding what you are looking for. I am glad that you didn't delegate that responsibility to the user.

- I think you should've put it on the AppStore first before sharing it; I would download it right away right now, as with many other people, but how many of us would remember Refresh a month later?

- I'd love to have branching history as well: so I have a tab and it has a linear history through which I can either go "forward" or "backward". Now, if I move backward in history, and click on a different link in that page, the history should branch at that point so that I won't loose the track of my previous "forward" pages when I click on a link in a previous page.

- Maybe not as satisfying, but I'd also consider developing add-ons for Chrome and Mozilla; some people would be reluctant to change their browser, but they'd be more comfortable adding exciting add-ons on a reliable browser.

All the best!

[0]: https://en.wikipedia.org/wiki/Tagsistant

> I think you should've put it on the AppStore first before sharing it

It's not a browser. It's a browser design. (plus a bunch of not-really-functional prototypes).

I hope it becomes reality but so far this is basically the The William[0] of browsers. The William also never happened even though we all really wanted it to.

[0] https://www.youtube.com/watch?v=2wBe3SDCcz8

The William is actually quite problematic from the UX point of view, most notably:

- There's a disconnect in space between the space being heated, the diagram of the heating patterns, and the heat control diagram. So you have three points of attention instead of, ideally, one, and traditionally, two.

- The whole interface has no undo. If I remove a heating space, reset clock, recalibrate weight, there is no going back.

- Removal of cookware deletes all associated data and programming.

- Touch interfaces while cooking are horrible and I hope they will never win.

Though, of course, the concept is from 2010 and the idea to add programmable heat curves (or, ideally, IFTTT of cooking) is a nice one.

Can you elaborate a bit on why you don't like touch interfaces while cooking?

Fingers tend to be dirty and slippery while cooking so tactile knobs are extremely nice, beyond the fact that you can use them without looking.

But in general, tactile knobs are always better than touch screen widgets if one looks beyond flexible function (tactile > non-tactile). Touch screens are more like an economy option when you want to fit in a bunch of functionality in a limited amount of space.

I seem to have too dry fingers most of the time, so can't get the touch controller to register. Not what you want if you need to lower the heat NOW.

In addition, a lot of the controllers seem to use the really silly plus/minus design to control heat, which means if you want to go from 8 to 3, you need to press 5 times instead of just one touch/twist. Fortunately my current cooking top has a bar so I can go directly, once it registers my finger that is...

I realize why they went with touch though.

Actually a lot of the ideas of The William are in real products now. Touch screens and timers are common nowadays and there are stoves that automatically reduce power before anything can boil and spill over.

Even the flexible zones are showing up in real products now (albeit with a lower resolution): https://m.youtube.com/watch?v=Jwljz2R0Z_M&t=54s

Induction hobs also have the similar property that only the pan is heated. If it is removed, the surface cools down even though the power is still on.

Yes that’s actually what these flex zone hobs use as well. They just spread loads of smaller inductors over a larger surface so it does not matter where you put the pots. The main difference is being able to control the power per pot rather than per predefined zone.

Holy hell! Never seen that before, and I very much want it now. Do you know why it didn't happen?

It was never meant to become a real product.

The concept was made by Greg Beck, a dude from Ohio, as some sort of a "hey, wouldn't it be nice if..." thing. Or maybe as part of his showreel as a product designer.

Sadly, his website[0] is down, but you can find a version without images in the archive[1].

The domain name is still registered though. You could write him an email and ask what happened. ;)

[0] http://becktothefuture.com/

[1] https://web.archive.org/web/20111208083212/http://www.beckto...

Thanks! I sent him an e-mail!

Iirc Electrolux was working on a fairly similar concept back in 2012-2013. Can't find the exact link for the touch screen one but these demos should be fairly similar:




Please do share his response, I'm curious (via response to this comment, or my contact info in my profile).

Maybe because creating a touch-sensitive surface that can heat up to cooking temperatures without melting the electronics next to it is harder than it seems?

Just because something is in an After Effects mockup doesn't mean it's practical to engineer.

Is it? Electric hotplates (and, in particular induction ones), already exist and support somewhat similar placement of user interface, even if of a more bare-bones one.

I suspect the heater grid might be tricky or expensive, though.

Even if the design was feasible, the product would undoubtedly target the high-end appliance market. In that space, however, you are competing with premium gas ranges rather than coil/induction.

> It's not a browser. It's a browser design. (plus a bunch of not-really-functional prototypes).

Oh, sorry!

The nerd in me appreciates the concept underpinning the numerous wee smart inductive heat zones, but as someone who prefers to cook with gas, the entire thing seems utterly misguided to me.

> I'd love to have branching history as well: so I have a tab and it has a linear history through which I can either go "forward" or "backward". Now, if I move backward in history, and click on a different link in that page, the history should branch at that point so that I won't loose the track of my previous "forward" pages when I click on a link in a previous page.

Implemented in Netsurf, with a nice history tree with screenshots :-)

I'd really like to see this in mainstream browsers.

Would be a real nice concept for the undo stack as well, UI challenges aside.

Glad to see the concept of work spaces realized quite thoroughly - I wrote about the need for such an approach before: https://rybakov.com/blog/open_tabs_are_cognitive_spaces/

I am intrigued, though, how well the browsing history concept scales up to dozens of tabs. In my experience, I only use the newest tab on mobile, while the other open tabs become some kind of living bookmarks.

Your article has a place in our reading list ;) https://refresh.study/reading-list

Oh wow, awesome! Glad my text was, in part, a source of inspiration :)

Somewhat coincidentally, I also used the term workspaces for a similar approach [0], and "bubbles" for what FF calls containers [1].

0 - https://cretz.github.io/doogie/guide/page-tree#workspaces 1 - https://cretz.github.io/doogie/guide/bubble

I kind of like your project just for the simple goal of not hiding complexity. As it is through hiding complexity that we get products whose minds we can not read.

As your approach is "Browser as IDE", have you thought of.. actually putting tabs next to tabs, inside of the main window?

Like Jupyter Notebook, but for websites?

After all, we all mostly do research and there is an action that springs out of our browsing activity. If that action is 'writing stuff down', why not do it directly in the browser?

The idea was less "browser as an IDE" and more "browser GUI like an IDE's GUI" meaning maximize organization and quick context switching in lieu of hiding everything. Granted IDEs often have tabs on top too, I just chose not to.

While I update it to newer Chromium versions regularly and use it as my main browser with great success, the lack of community interest in the project limits my desire to put much work in to new features unless I really want them.

This article and the idea of tab history solidify an approach that I have always taken to browsing that I've dubbed "immutable browsing". I ctrl+click every link to make a child and only purge the roots of the tree to roll up the children when the roots habe no value to me anymore. On a Wikipedia or Github project spelunking expedition, I can get many levels deep, sometimes ctrl+clicking dozens of pages as background child pages from a single page.

This is correct. Browsers need to become “spacers”.

The spaces need to be publishable and moderateable and multiuser and all the other things we expect for web objects though.

Nice progress made on this project.

You should check out Workona's workspaces. We've built some of what you've mentioned as a Chrome extension backed by a web app.

Our workspaces (persistent windows with names) remember your tabs automatically, are multiuser/collaborative, and even have a URL (like any good web object). You can also access your workspaces on multiple computers, so you can start a project at work and finish it at home.

We're officially launching out of beta next month, but you can sign up now if you're interested:


Just tried it -- this is a great idea. Unfortunately, I can't use it because Workona spaces don't "containerize" accounts I'm logged into.

For example, I need multiple spaces where each space is "dedicated" to a single Google account -- it has multiple tabs open that are logged into that particular Google account.

I tried to create that setup, but the accounts "leak" between spaces: logging into a Google account in one space doesn't allow tabs in other spaces to be logged into a different Google account.

If you ever implement proper account isolation per space, please notify me: I'd like to be a beta tester.

Glad you like the concept. You're correct that we don't attempt to do account isolation as that isn't the purpose of our workspaces. Our workspaces are meant to be used at the project/meeting/workflow level, not at the Example Inc./Acme Inc./Home level.

As another person suggested, you can use Chrome profiles with Workona installed on each, but that likely isn't what you're after.

> For example, I need multiple spaces where each space is "dedicated" to a single Google account

Could you sign in to chrome in your different google accounts, with an instance of workona per google account? It's not perfect, but might do what you want it to do?

This looks promising. I've tried to adjust some of my workflow around a concept like this using Spaces/Task Switching apps but eventually just fall back to old habits. Will give it a try.

Glad you guys are launching! I tried workona for a week, but couldn't justify switching away from firefox. It seems though, you added a lot of features, maybe it's time to try again..

Thanks! We'll add Firefox support soon. In fact, they have some additional APIs that we're pretty excited about.

Does Workona save tabs and their state, or only the URL?

For example, if I'm halfway through filling out a form but switch workspaces, will it lose my progress? Can I have multiple workspaces open in different windows?

It depends. If you are just opening the workspace for the first time, then we will just be loading up the URLs. We do preserve tab state (form data, page scrolled, text selected, etc.) when you switch between workspaces that are already open. That way the tabs don't have to reload as you repeatedly switch contexts.

By default Workona opens workspaces in the same window, but you can command-click (Mac) or control-click (Windows) to open a workspace in a new window.

This video will probably help:


I love the concept, but...

Ever since the early Internet; pre Firefox era, I have been a heavy Tab user. I am possibly on the higher end of spectrum, my Safari currently has 410 Tabs opened. It sort of expands and shrink during the weekend. In some days they could blow up to 500+ Tabs, more like a reading list.

Quite a few of them are youtube links that are acting like a mini playlist or file selection of music. So having a Video as group selection shown in this concept is a godsend. I have been wondering why Safari Multi Tab page doesn't have this. ( And Safari Tab Overview is kind of stupid, it reloads / reactive all the tab you have when you press overview, making huge memory and CPU spike as well as jack )

But over the years I have came to the conclusion, most of the consumers don't use Tab like us. That is why those mini Tabs in Chrome that shrinks to ridiculously small size don't annoy them. They don't check email on the web. Nor do they do much comparison. Their web browsing flow is very linear. While my web browsing flow is extremely jumpy. Reading on something could bring up dozens of other questions in my head.

So I am wondering if there is enough market for this browser? Don't get me wrong I love innovation in the browser UI space and I love the new History timeline. But I am slightly worry about its sustainability.

There's also people like me, who has in the past used tree style tabs and other ways to manage their crazy number of tabs while browsing, but since then realised that it's not for them.

I used tabs like they were bookmarks, so I started using bookmarks more. I used tabs to keep track of history, so I started using the history more. I used tabs to do things the web browser shouldn't do, such as check email or play media, so I started doing those things in the right application (email client and media player, respectively).

My tab use dropped immediately.

I now have a plugin that prevents me from opening more than n tabs (in my case, I have set n=7 right now), which forces me to linearize my naturally sprawly browsing. I need to either finish one task before going on to the next -- or make a proper core dump of it, which includes mental notes as well as links to browsing session. This has actually helped me pick it up faster later on!

It has also prevented me from stressing about unimportant things because they are open in the web browser, telling me that they are left unfinished. Now I can prioritise the important stuff easier.

Yes, bookmark and history are exactly what the function of those huge number of tabs. Along with reading list.

Example, I now have a list of International shipping website as well as import tax web site that will likely takes me hours if not days to finish. I cant finish them now and move on to next thing. I will have leave it there, and make it like a to do list.

To support your point, data gathered from the Firefox user population a few years ago showed that, IIRC, the median number of tabs was somewhere around 2. And a lot (I don't remember the proportion but it was unexpectedly large) of browser sessions lasted a couple minutes. As if lots of people opened links from outside the browser, read it, and closed the browser shortly after.

Comments like this remind me how out of touch I am with the average user.

This sounds like what I sometimes open Chrome up for - to try and see if something will work in a different browser, and then closing it back down and going back to Firefox.

The tyranny of the minimum viable user


Tabbed browsing: a lousy band-aid over poor browser document and state management


What if the Web was filesystem-accessible?


>Their web browsing flow is very linear. While my web browsing flow is extremely jumpy.

I'm curious what you mean by "jumpy." The linearity I understand, and notice it when I'm researching with coworkers or shoulder surfing with someone.

I usually us saved links or bookmarks to preserve state or come back to things later. My primary use of tabs is to approach research as a sort of branching tree. Start with a page, any interesting link in it gets opened in a tab. I follow each branch in turn until I've got a handle on whatever I need to. Similarly, I use the tabs for comparison shopping, so they're useful for "shortlisting" items of interest.

how do you manage so many tabs? I struggle to find my active tab when i have 10 open. Maybe I should change my theme.

Here’s an interesting question to turn over this lovely Sunday afternoon.

What stops you from implementing this yourself? Why?

Chrome is supposedly open source. So is Firefox.

It’s an important question, because the answers lead to the reasons why we’re subservient to megacorps. And it also shows merely being open source and free to use isn’t enough to ensure freedom.

We tend to believe that (a) browsers are an open ecosystem, yet (b) we can’t modify them. This is a contradiction.

It would be a massive piece of work, is most of the answer.

People generally don't want to pay for software, is the other part of the answer.

Building Chrome requires a supercomputer and not everyone has one around.

Why is that so? Because there are enough crazy people who cannot stop writing new web standards. For example, things like Shadow DOM and web components, that are not really necessary and can be implemented using existing specs.

>Building Chrome requires a supercomputer and not everyone has one around.



Chromium comes in at 120 SBU, which while large isn't impossible. That comes to about 3-4 hours on my laptop.

Building software is not some esoteric pursuit only high priests can perform. It's good practice for software you're familiar with, require performance from and esoteric options most people don't care about.


Ninja, by default, sets the number of jobs to the number of available threads. If your machine is memory constrained, with regard to number of threads, you should reduce the number of jobs by passing the -j# switch to the ninja command below. Plan for roughly 1.25GB per thread while linking WebKit to avoid out of memory errors.

My laptop is memory constrained, I have yet to build Chromuium from source. How much RAM on your laptop? I've been looking for a laptop with 32GB RAM.

>My laptop is memory constrained

you should reduce the number of jobs by passing the -j# switch to the ninja command below.

I have had builds of linpack back in the day taking 24h+ on tiny systems. Again, building software isn't rocket science and you don't need super computers for it.

What's an SBU? (I tried to search but only found results for Stony Brook University)


unit for compilation work relative to a reference package.

What makes you say you can't modify them? Are you referring to modifying your copy, or are you referring to modifying the main branch and thus other people's copies? You're trying to blame this on "megacorps". However, as a user, I appreciate the maintainers at Mozilla for keeping Firefox well-tested and well-maintained, part of which is scrutinizing feature proposal and PRs/patches to make sure that their end users would want them. This is such a strange criticism of Chromium/Firefox. At least you can modify your copy at all.

Your other comment suggests trying to add a new web standard but IMO the unending stream of web standards is why only "megacorps" maintain browser engines, because only they have the resources and care enough.

Either way, it's a logical leap to suggest you need to modify Chromium or Firefox to implement this in the first place. The concept from TFA can be implemented with webviews without seeing any code from any browser. To answer your first question, the only thing stopping any implementation is time.

Because open source monoliths is impossible to maintain unless you can dedicate most of your life to that. Split that to many small individual, replaceable and easily maintainable pieces first.


> We tend to believe that (a) browsers are an open ecosystem, yet (b) we can’t modify them. This is a contradiction.

I think the idea of it being "open" is that you can do whatever you want within the boundaries of the sandbox you are provided, right? Closed ecosystems can restrict what you can do beyond the APIs provided (no VPNs, no porn, etc.)

Even the “no porn” is giving up freedoms.

The megacorps wouldn’t have allowed the web to evolve into such a massive sandbox if they’d known how much of an effect it would have. Just like Apple didn’t, with the App Store.

The same situation has already happened with Chrome, but with browsers themselves. Try submitting a PR to chromium that implements a new web standard and see how far it gets, let alone a change to the UI.

All of that begs the question: why is it like this? Theoretically there’s nothing stopping you from writing a browser in JS and delivering it as a website. Then users could customize it however they wish. And if you deliver it as a standalone executable, you can do things like add native torrenting support.

> The same situation has already happened with Chrome, but with browsers themselves. Try submitting a PR to chromium that implements a new web standard and see how far it gets, let alone a change to the UI.

FWIW, there have been various new features/standards implemented by people from various companies (some people will have heard of, like Samsung and LG, and others you probably haven't, like Igalia) in Chromium. (UI changes I know much less about, given I rarely look outside of the Blink bubble.)

> Try submitting a PR to chromium that implements a new web standard

Please don't. There is already oversupply of web standards. Try writing a new browser engine from scratch; it would take ages.

You can modify them, and a lot of the readers of this site have! The problem is convincing everyone to care. Or even one person. This is the same problem all founders struggle with all day long. Obviously, many are successful.

Business gets in the way of a lot of things, I'd say. "No distractions" mode here would strip the branding, the ads, the engagement tricks utilized by websites...

It'd be fun to see this as an OSS project, though.

Firefox already has a so-called "reader view", which basically strips away the distractions. It would be nice to see this implemented more thoroughly, so navigation menus are retained and become standardized/clutter-free.

And it gets around some adblock blocks :)

Yeah, I think we need better collaborative filtering and/or ML for ad-blocking. And the browser should always make the website believe that the ads are shown (no altering of the DOM, but instead alter a copy of the DOM).

Yes, but one advantage of blocking ads is not having to get the content. On a slow uplink, such as Tor, that's a major issue.

True. I guess you could load the content from a client running at a central service, and send the stripped-down version to the final client. Of course, you'd have to trust the service.

I think hanlon's razor applies here. The Web is ridiculously complicated and it takes a supercomputer to compile a browser capable of running The Web at full speed. Not to mention that people have done it -- see, e.g., qutebrowser or next-browser.

It's called the unix philosophy and anything that touches the web is far, far from it.

I want to see something like https://web.hypothes.is made native in browsers, so you can have community overlays on websites and commenting based on a group you subscribe to, automate common operations teams perform on web pages and leave notes for them etc. with out the websites itself being entirely in control of the entire ux and functionality.

We actually thought about adding a similar function in the beginning, but discarded it later since we wanted to focus on the existing key problems, rather than adding productivity features for power users. Nevertheless it is an interesting topic worth exploring.

In the same vein it would be interesting to explore mechanisms of automatic discovery - e.g. proposing tabs that have close horizontal proximity in other people's collections. Sure, it opens up the questions of privacy and data collection, but it could provide a much needed alternative to using search engines for discovery.

I think the vision is more that it’s a meta web rather than a power user feature. Glad it’s on your radar

I'd like that too. And perhaps the browser should know about the concept of "user-reviews" (where the reviews are customized per-user, i.e. computed based on the groups you are in, and who you trust, etc.) Thinking about it more, that's what the web needs most: reviews that we can actually trust.

Do you remember Google Sidewiki? https://www.youtube.com/watch?v=CsjJOsx84MA

Haven’t seen that before, my thought is that it’d be more like a close knit group and also have workflows.

Didn't Genius do something like this? I remember seeing random political articles with annotations from pundits a while back.

Moving back in time was the most interesting new idea in the preso to me. Makes what you do in the browser quite accessible. Between this and web archival it might become possible to have easy, on-demand access to just about any personal web history. Imagine being able to travel back in time 10 years and see the websites you perused.

>Moving back in time was the most interesting new idea in the preso to me. Makes what you do in the browser quite accessible. Between this and web archival it might become possible to have easy, on-demand access to just about any personal web history. Imagine being able to travel back in time 10 years and see the websites you perused.

That was a killer feature as presented. In practice it might wind up a little iffy though. Web links aren't static, so the "history" you see might not actually be the pages you were reading. Especially for landing pages for news sites.

I can also imagine it becoming a memory hog for people with certain types of browsing habits. My wife, for example, basically never closes tabs. She just lets them purge on occasions when she needs to restart, which is basically just for OS updates.

Imagine giving every new browsing history tracker the ability to see what websites you looked at 10 years ago.

Maybe it's the curmudgeon in me, but while I salute the ingenuity in this UI/UX idea, I think we need a new kind of browser engine at least as much as a new kind of browser interface, perhaps more.

I used to worry that WebKit would eat the world, for awhile Trident did eat the world, and now I worry about Blink.

Maybe you should take a look at Servo from Mozilla: https://servo.org/

A few things from Servo were integrated in Firefox 57. But the project is far from complete I think.

I'm aware of Servo. Servo is neat to play with, though it's pretty bad with older legacy HTML pages. Also, I'm not sure how likely Mozilla is to develop it into a product; it seems more like a testbed for later Gecko evolution (which, to be sure, is still a good thing).

> I'm not sure how likely Mozilla is to develop it into a product

It's unlikely to become a standard desktop browser, but the roadmap [1] does mention investigating options like a standalone mobile browser [2] or perhaps to be used embedded in an app, e.g. for packaged web apps.

[1] https://github.com/servo/servo/wiki/Roadmap [2] https://github.com/servo/servo/issues/20855

Firefox already includes parts of Servo, so it’s not a question of “how likely” it is to be used in production. https://en.wikipedia.org/wiki/Quantum_(Mozilla)

> so it’s not a question of “how likely” it is to be used in production.

I believe you misunderstood the comment unless it was changed. The commenter questioned the likelihood of it being developed into a product and you responded as of they had questioned the likelihood of any of it being used in production.

Fwiw, I agree with the commenter and am unsure how likely Mozilla is to develop it into a product. They do seem to incorporate pieces though.

I think you and the commenter share some misunderstanding. Could you distinguish what you mean by developing Servo into a product vs. “incorporating pieces”?

Are you saying that Mozilla eventually replacing Gecko’s style system, DOM, compositor, and networking system with Servo’s (as the article I linked describes) is just incorporating “pieces” of Servo?

The product that Mozilla is developing Servo into is Firefox. I don’t know what you expect—for Mozilla to replace Gecko with Servo all at once? For Mozilla to do the same thing they are doing now but just give it a new name?

> Could you distinguish what you mean by developing Servo into a product vs. “incorporating pieces”?

Do you remember seeing screenshots of "servo"? Running a Servo executable (or browserhtml or whatever) is/was a thing. It was branded as being developed as an alternative browser to FF. It was even using cannibalizing FF pieces (e.g. Spider Monkey) not the other way around.

> The product that Mozilla is developing Servo into is Firefox. I don’t know what you expect—for Mozilla to replace Gecko with Servo all at once? For Mozilla to do the same thing they are doing now but just give it a new name?

That is the case now, though they have names for the pieces, the name "Servo" still seems to be the name of the full browser (coopted by the VR group IIRC). It was what everyone expected to not necessarily replace FF all at once, but be developed alongside as an alternative. This history was clear, the name Servo as a whole browser was clear, and the misunderstanding is pretending that was not the case.

Not only is your understanding of history at odds with the facts, which anyone can look up for themselves, but I am extremely confused at why the ill-defined concept of “Servo as a whole browser” is important at all.

Servo is a browser engine, not some VR project (where the heck did you get that? Go to https://github.com/servo/servo ); what’s more, Firefox is “Servo as a whole browser,” whatever that is supposed to mean. Of course you are able to run Servo “by itself,” how the heck do you expect Servo developers to be able to test their browser engine?

I’m not sure what weird dramatization of history you are trying to create, but it’s all the more bizarre in that (1) this all happened in the open, so you should easily be able to see what you are saying is and was simply not the case, and in that (2) far from Servo being killed, it is the future of Firefox, so what on earth is there to be upset about? This has been surreal; good luck.

That's not what I meant. I meant whether Servo the engine is likely to be a standalone product (on its own merits along with Gecko).

I think you misunderstand what a browser engine is, then—read the Wikipedia article I linked and tell me how the “stand-alone product” that Servo is going to be is not Firefox itself. Massive parts of Gecko have already been simply deleted from the codebase and replaced with components from Servo.

“Firefox” is not a browser engine, it is a browser UI. Gecko and Servo are browser engines.

The only thing I worry about is high complexity. If its so complicated, that only a few megacorps can implement it, it should be abolished. That would especially mean to scrap most web standard concepts and replace it with more general concepts.

Pretty nice. I use a Safari add-in called Sessions to create sets of browser tabs I can recall later. Super handy when you are researching something new or just want to persevere your current multi-tab browsing session for later.



>Sessions will not install on Safari 12 or later, as Apple has officially deprecated Safari Extensions.

Which is being replaced by Safari App Extension Platform. Did I read it correctly that App Extensions can only be downloaded from the Mac App Store?


Safari Extensions from the Safari Extension Gallery still work, but standalone ones don't.

Looks like Tab Groups or Tab Trees, just 10 Years later.

That's exactly what I thought until I saw the proposed timeline mechanism for exploring session history. That feature looks very useful.

It looked entreating but i still can't come up with a use case; I'm very very rarely browse through my history. I only search for a specific item

Trip planning comes to mind. If you have a bunch of booking sites, along with hotel information open, then you switch contexts to something like airfare booking. It's conceivable you want to eventually come back to the hotels if you haven't finished booking, and getting all X tabs open again would be helpful in that case so you don't have to individually remember all the hotels you were looking at.

If only sessions on travel-related websites didn't expire already after half an hour.

That's a good point, although some will re-search once the session is expired. It still helps with reference sites (like the hotel's/airline's/cruise's/etc page).

Ever since Opera came out with tabs I've been using huge session files. Back in Opera 5/6 that meant maybe 50-100 tabs. These days my session files are more towards 1000 tabs.

I save my session file with semantic naming every week or so. I would love to have a graphical timeline of all tabs like this to explore older session files as well as my current one.

I search for a specific item in my history all the time, but it can sometimes be difficult to find what I'm looking for via Chrome's itemized list search, especially when the <title> tags are all the same. I think this would help me find it, in relation to my other activity and time.

I'd love to see this built out - I was disappointed there was no download link.

Reminds me of Firefox Container Tabs and other advanced ffeatures

I hope they build this!

Making web browser software can be very profitable if you get a large userbase. Mozilla had $520 million of revenue in 2016!

More browser innovation would be great, considering how many human hours are spent browsing the web. Small browser level increases in efficiency with better tools for web browsing could have massive impact.

And, building browser companies can be a path to making lots of money, so it's not a charity.

On desktop chrome you can press ctrl+f find a link on a page using search then press escape and then press enter to follow the link without a mouse

I wish any browser on iOS could allow this, I use an iPad and I can shortcut myself to anything on safari with my hardware keyboard, except clicking links, it's really jarring to jab the screen for that part

Really need something like Vimari/Vimium for iOS Safari with hardware keyboard. I wonder if there are any custom browsers that support that.

Looks really cool. I wish there were more of these types of fully fleshed out design thesis. A couple years ago someone did one on Windows 8 and it had an outsized effect on the development of windows 10, someone had a similar idea for office, but I don't know how much of an effect it had. either way it was fun to read and look over.

Some of those features can be done today with any browser, such as spaces.

Just create a virtual desktop for each space.

It really does save a lot of time in the long run. The added benefits of virtual desktops is you can also recall more than just browser tabs.

On Windows 10, all you have to do is hit CTRL + Win Key + Right arrow to launch a new desktop.

This is entirely different concept and has no implementation on mobile which looks like to be a target platform.

Interesting concept. Looks like a bunch of ideas from other extensions and apps that are already doing these things.

Tab management: Pin tabs seems to do this already.

Saving & recalling: Tons of "spaces" extensions already exist.

I'm not a big fan of the netflix layout.

Smart link previews: Nice to have an image preview but the auto categorization of links would probably not work so well. So this would be probably only be used for main well known sites.

Browsing History: That timeline actually looks really nice. Not sure how well it would work, but that's actually innovative for a browser.

Tagging: So... a bookmarking system?

Focus on whats important: Nope nope nope. I don't want my browser to become an app that strips out what makes a website unique. Most often this functionality just breaks the UI or strips away code that breaks the functionality.

> Focus on whats important

Firefox has such feature (reader's view) and it works surprisingly well. It really helps, especially when the text is difficult to read (contrast issues, small fonts, etc) or when a page contains lots of elements that try to distract you from main content.

Just wanted to point out that Chrome has Chrome Profiles which can be used for "work" "thailand" etc spaces.

Maybe we can extend this with a better UI + ability to move pages between profiles (this creates some serious security problems though)

Well, I think most of the concepts are design-centric.

We have history, tagging and bookmarks but I barely use them in modern browsers.

This UI makes it more orthogonal (it seems).

Firefox has tab containers which are similar but can be open at the same time on the same window.

That's a browser I would love to use on my desktop. Two thumbs up!

FWIW, you can get similar "Spaces" functionality with Toby extension: https://www.gettoby.com/

Nice idea, I have definitely felt that the browser market space has become to stagnant and bland. I am surprised you are using cookies for what mostly seems to be a product outline.

Wow, looks great, good job. I really wish this was available.

Yeah, I so want this. Funny how these features take so long to arrive then seems obvious afterwards.

I made it 3/4 of the way through the video, then clicked around the website in frustration looking for the download button.

This reminds me a lot of Firefox tab groups/panorama which was removed a few years ago. I like the history idea, there are some really good ideas there. I tried finding something in my history on Firefox recently and it was an awful experience, so maybe someone will be inspired by this.

> This reminds me a lot of Firefox tab groups/panorama which was removed a few years ago

It's back: https://addons.mozilla.org/en-US/firefox/addon/panorama-view...

I tried that addon, and it is not the same thing. Panorama was an overlay, but this opens a new tab to view your groups. This doesn't really work when you have hundreds of tabs. I'm pretty sure that the old Panorama view is not possible with webextensions, although I haven't really looked into it.

I don't get what the difference is, except that in the old version there was no tab bar in the overview, whereas in the new version the tab bar is still there with just a single tab that says "Panorama view"? I don't see how that interferes with the number of tabs you have? I have quite a few tabs as well, no problem there.

This is what it looks like if you press Ctrl+Shift+F: https://addons.cdn.mozilla.net/user-media/previews/full/200/...

If you then click one of the tabs, the tab bar gets populated by the tabs in that group.

(Perhaps you tried this extension before the tab hiding Web Extension API arrived? That only arrived in the latest version of Firefox, IIRC.)

>Perhaps you tried this before tab hiding

I might have done so. The problem I had is that I was in the middle of a few hundred tabs and opening panorama opened a new tab at the end of the list making me lose my place. If it hides all tabs when you open panorama it might work well enough.

I'll try it again, thanks for letting me know.

In that case I think you're going to like the extension now - I did :)

>> And while the web itself has undergone big changes in the last twenty years, browsers have remained largely the same.

Yeah, it's largely the same : we open app, enter the URL and it shows a web page..But cmon, stating that IE6 is largely the same as Chrome or Brave is kinda exagreggation.

Containers is available in Firefox, however the usability and ease of usage should be something like shown in video. Its pretty awesome to shift between containers.

Timeline seems like barrowed from Windows. This will be another awesome feature I love to see

I used to use containers all the time (tab groups per activity/mind state is a thing I've always craved for) But they're buggy. Really buggy. Twice I've lost a trace of all the containers and tabs that were open (over 40).

Kinda confusing TBH. Browsing should be simple, why convolute it with the Productivity App ideas?

Great design and media, got me interested!

Interesting triplet of Chrome extensions.

It's like Netflix for tabs

Did this has tabs on the side?

I'd use it.

Disappointing that a website about making the web better still wants to track me with cookies for god knows what reason (also non-GDPR compliant despite them being based in Germany).

They clearly state that they use cookies for analytics. They are not collecting PII. Their Google Analytics implementation uses `anonymizeIp`.

I would not trust a company whose bottom line is proportional to how much personal data they collect to just honour an "anonymise" argument.

Not to mention, analytics still wouldn't require a cookie unless they want to know which other sites I visited, which is none of their business.

The cookies are also used to distinguish visitors. I'd recommend using private browsing or a GA blocker plugin if you're not comfortable with that.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact