Multi-Account Containers

lucideer

@pafflick said in Multi-Account Containers:

Some features (like this one) might have hundreds of votes, but due to their nature, they will be used only by a small percentage of users.

A few people are taking this at face value, but I think it needs some interrogation as I don't think there's really any evidence for (or against) this assumption. It is just an assumption, a potentially very inaccurate one.

Firstly, assess general population interest: people are interested in features they're familiar with I want X which I used before in <other place>, so you need a comparator. Firefox's MAC has exceptionally poor UX, so the low interest in the general population is going to be affected by that. So interest within this demographic can't be reliably measured.

Secondly, Vivaldi's current user-base: this user-base doesn't represent the average web-browser user, the average Vivaldi user will tend to be more engaged in their web-browser's features, so interest may be much higher in any given technical feature than in the general population. This demographic is the most measurable - obviously this is not only the top-voted feature request, but 3 of the top-10 are session-related - but even with that you're still relying heavily on people's experience with a feature that's never been done well yet. Good UX is important to ensure people use any feature.

Thirdly, Vivaldi's target audience: this is not the same as the current user base as there are thousands of people who want to use Vivaldi but have refrained for various reasons: resistance to closed-source, resistance to Blink engine, lack of <key feature> (like MAC - my own personal blocker), etc. etc. These users are obviously impossible to measure.

Conclusion: You cannot dismiss the possibility that interest in this feature could be very high.

To further rebut the argument that demand is needed for such a feature, Vivaldi just introduced Workspaces, a really cool & innovative feature that noone asked for. Noone asked for it because it's novel: noone has any prior experience using this feature, so they can't know they want it yet.

Existing demand is not a requirement for future usage & demand.

The privacy benefits of an MAC-like feature with good UX (which has never been done) should speak for themselves.

lucideer

Follow-up comment to clarify, just in case my above comment comes across as ignorant/inconsiderate:

I'm aware the feature is difficult, and that this means it may never get implemented for that reason alone - that's understandable & reasonable
I'm aware the feature is dependent on a third-party (Google) accepting changes to an upstream dependency (either Blink or Chromium, I'm not sure which, possibly both), and that Google's acceptance criteria may be a blocker to this ever being implemented - that's understandable and reasonable

I just wanted to make an argument that if it's possible and semi-feasible, I would hate to see it de-prioritised due to the perception of demand not matching investment.

royalkin

The lack of support for multiple account containers or something similar is the one feature that is preventing me from coming back to Vivaldi. The feature must be baked into the browser as all such third-party extensions either don't work or don't work very well.

This is a must-have feature for me; it's a work requirement. Therefore I'm stuck using Firefox with its poor customization features.

It's sad that this is still not in Vivaldi. VIVALDI TEAM, PLEASE IMPLEMENT A MULTIPLE ACCOUNT CONTAINERS FEATURE!

stardepp

@royalkin Try this: https://vivaldi.com/de/features/workspaces/ or: https://help.vivaldi.com/desktop/tools/user-profiles/

royalkin

@stardepp It doesn't look like either of those allow me to have multiple tabs with different accounts logged into the same website/service. That's what I actually need.

Ayespy

@royalkin Best info I have so far is that the Blink/Chromium architecture does not allow this. A major patch/rewrite would need to be done and then carried forward (re-added) in every subsequent build thereafter. Not saying this is impossible, but it's not merely a matter of "want to" for the crew.

lucideer

@Ayespy said in Multi-Account Containers:

and then carried forward (re-added) in every subsequent build thereafter.

Or upstreamed. But that would depend on Google being OK with it...

Ayespy

@lucideer said in Multi-Account Containers:

that would depend on Google being OK with it...

That it would.

IchKannEchtNix

@stardepp You do not understand how container tabs are working.

Open a Google.com tab. Google Mail. Google Calendar. Whatever. Login. There will be cookies created with information about your account.

Open a second tab with a random web site which uses Google Ads. Google Ads can on this web site can now access the Google.com cookies, and save the information that YOU visited this Web Site.

Container Tabs:

Open Google.com and login as above. WIthin a GOOGLE container. Cookies will be created.

Open any web site which uses Google Ads in a different container. Google Ads cannot see the cookies from Google.com and you are a different person for Google.

Container Tabs allow to separate content of web sites: each container has its own cache, his own cookies. Within one browser session, you can open Amazon.com, login and manage you account. With a different session in a different container, you can search on Amazon.com (without login) for pink toys for you daugther. Amazon.com can not merge the pink toys searches with your account, because the cannot see the cookies.

You could say "Thats what private browsing is made for". The difference is: private browsing does not save cookies for later uses, but container tabs to.

In combination with extensions like fingerprint defender, you can surf within the web without Google knows where you are: open Google web sites like YouTube and Google.com in a special Google Container, and you can visit all other web sites in different containers without that Google Ads can access these cookies and link the information with your profile. And because fingerprint defender add ons create a unique finger print each time, even that fingerprint does not help them to identifiy you.

To get the same feature in chrome engine based browsers, you have to open each single web site within a private browser window. A private browser window for google, for amazon, for facebook, for twitter, for vivaldi. Five different windows! And when you close them, all cookies are gone, and next time you have to log in on each web site again. I have usually 200+ tabs open. Should I create dozens of browser windows?

In Firefox with Container Tabs I get all this within a single browser window with the bonus that the cookies are stored for each single container, and the next time I dont have to login again.

Thats the only reason, why I use Firefox instead of Vivaldi. Stacked tabs, page tiling: I love these features and I want them in Firefox. But Container Tabs are a HUGE privacy improvement. I removed Vivaldi and come back from time to time to check if this feature is available.

And I dont get it, that Vivaldi implement useless toys like Phillips Hue integration, but does not implement Container Tabs. Not even an official statement exists why this feature is not implemented. Technical reasons perhaps? I can understand that, but than I am a lifetime Firefox user.

Privacy (Container Tabs) > Quality of Life Features (Tab stacking, Page Tiling, mouse gestures)

Workspaces and User Profiles are NOT a replacement for container tabs.

pauloaguia

@IchKannEchtNix said in Multi-Account Containers:

Not even an official statement exists why this feature is not implemented.

https://forum.vivaldi.net/topic/25289/multi-account-containers/214

TheAMan006

Do Workspaces have seperate cookies?

Ayespy

@TheAMan006 Not so far as I can tell. That would be containers, which is a structure that Chromium browsers do not, in any way, support other than the option to use multiple profiles.

becm

@TheAMan006 appart from profles, Chromium differentiates storage scopes between normal and private mode only.

Not an easy way to work around (also why private tabs are not easily doable, scope separation is done on window type level).

But this can (in some cases) suffice as an alternative/2nd login container.

TheAMan006

I guess not basing it off of Chromium gives Firefox a lot of competitive advantages, as in, unique features that others in this landscape simply can not offer! I reckon it has Multi-Account Containers, and Tree Style Tabs too!

Ayespy

@TheAMan006 No, Firefox does not have treestyle tabs. It (last I knew) does have an extension that shows them.

Vivaldi does have the Windows panel that shows tabs in treestyle. It's built-in, not an extension.

TheAMan006

This post is deleted!

wulreuben

This feature is a selling point for a lot of freemium browser environments like Wavebox, SessionBox, Ghost.

On the topic of extension solutions, SessionBox's extension solution is sub-par but still has 300k users, this doesn't seem like a niche.

If sustainable business models are being supported in part by this feature as a selling point, I'd say multi-account containers are worth reconsidering for pipeline.

Of course I'm not clued in on the internal considerations for the Vivaldi team, but from the outside this looks like an opportunity to get ahead of the curve with work-from-home and the likes on the rise recently.

wulreuben

I have no programming experience but from what I've been reading elsewhere I have what I think would be the process to make this happen. Does seem like it would take a lot of work tho. Just wanted to make it happen in my mind at least

make new url schema made to handle "containers"
-- e.g. "vivaldi://container/[container type]/[container name]/..."
use First-Party Sets (FPS) or Cookies Having Independent Partitioned State (CHIPS) or similar to hold cookie partitions inside the new container urls
-- users would turn workspaces, profiles, etc into containers as a menu option or something and they would become urls like the example above
-- could add option to create container with or without current cookies, settings, etc.
change browser settings from being stored as preference files to cookies where needed
-- local urls would request cookies which would be partitioned to create different setups inside different containers. importantly works offline (i think)
modify browser UI to be populated completely with embed elements;
-- panel, web views, address bar, status bars, icons, everything.
make/keep persistent elements available for things like mail and calendar vivaldi profile, menu with unpartitioned cookies
make persistent elements for functions necessary to switch between containers in the form of:
-- profiles, workspaces, sessions, tab stacks, etc.
-- populate these with a global list of top-level container urls in the appropriate menus as well as any embeded items in the currently open container
-- (so for example, persistent workspace menu shows all container-ized workspaces, but only the embeded workspaces assigned to the container you're currently inside of, as these are the only ones the container knows exist)

Result would be switching profiles, workspaces, etc into partitioned settings, cookies, logins, etc. all within the same browser window and under the same Vivaldi account.

Side panels, address bars, tab settings, bookmarks etc. would populate with items specific to the containers partitioned cookies.

Tell me if I'm way off base with this or not, but I thought I'd lay out how I saw it working in my mind... Just for fun hehe

caiocgo

I would love to see this implemented on Vivaldi! I need to use multiple accounts of a given service for work purposes and this would come really handy!

bdmendes

I would love to see this tightly integrated with the new workspaces, with a simple 'isolate workspace data' sort-of option. This would mean I would always have my personal email logged into my Personal workspace, and my work one on the respective workspace. This would actually make multi account containers useful and organized, unlike in Firefox imo.