Vivaldi is developed by a team created by the former Opera founder and CEO. Vivaldi kind of tries to re-create the spirit of Opera, but I think it's going to be too hard to do it.
For those who didn't use it, Opera was first a paid browser (which limited its reach), then an ad-supported browser (which again limited its reach). It then finally became a free browser but by then it was too late.
It had its own, super fast, rendering engine (I forget its name, Presto?). It had a built in email client, feed reader, calendar (unfortunately with no Exchange or Gmail integration), a notes app, a powerful download manager and even a Bitorrent client. And a TON of features and UI flexibility.
It was super compact, a marvel of engineering and UX design that managed to pack all those things in a package of about 5MB at the time, and you wouldn't even see or load the extra functionality like the email client if you didn't use it.
Unfortunately with HTML5 and the Chrome-ification of the web, it couldn't keep up :-(
Vivaldi tries to do the same on top of web techs and web techs just can't handle it. Web techs are almost as flexible but they're really slow and bulky.
> Vivaldi kind of tries to re-create the spirit of Opera, but I think it's going to be too hard to do it.
Oh, it's definitely recreating the experience of Opera. You encounter an annoying bug, you report it to their private bug tracker, and then they'll fix it in two years. Maybe. If you're lucky. And you don't get to know its status in the meantime. And the macOS version still feels like an afterthought sometimes.
Still better than Chrome and by far better than Safari.
Seems pretty quick, at least compared to Chromium / Firefox ;)
> you don't get to know its status in the meantime
Toward the end of Opera's life (pre v15) either an employee or a member of the community (can't remember which) did actually set up a public bug tracker tracking the status of publicly-reported bugs. It didn't show comments or anything like that: just the original bug description, report date & status. So that was... something at least.
They did also give private bug tracker access to a subset of community members, by invite.
Yeah you can still ask people who have access to look up the status of a bug for you[1]. It just feels so needlessly convoluted. Since they're already using Jira, it shouldn't be too hard to set it up with proper access controls so you could see your own tickets, at least. They must have some reason to not do this.
If memory serves, they removed the user agent because one of the Google properties was serving it substandard HTML/JS even though they use the Chromium engine.
Think about it for a second. They removed the very most important metric for a product selling content leads and placement (search bar default bookmarks, placement on the start page) because _one_ website had a bug?
I dunno, I've submitted four bugs. One solved within a day, one after 20 days, and two still unresolved as far as I know. It could of course be better, I don't think it's quite like you describe.
I used it exclusively for about a year (iirc) on Linux and had only some minor crashes. When they switched to Chromium I switched to Firefox and never had a single crash since. (It's not only luck, I disable all the feature-creep I know of in about:config)
I vividly remember how it took them several months to make Opera display Flash content correctly (at 2x scale instead of a blurry upscaled-after-rasterization mess) on retina displays when I just bought my macbook. Other browsers had it done in a matter of weeks, days if you installed a beta build. And that was 2012, back when flash was an essential part of the web.
Flash was not an essential part of the web in 2012. Maybe, if you had some weird work tool you had to use, some video players on the web but that was mostly it. In 2012 flash for Android was killed and it was very well known that it was going out of the window in the near future.
Many video players were still using flash. VKontakte still had its flash app/game platform going strong, though you did have the option of using an iframe that loaded a web app from your server instead. Flash support was only discontinued there around several years ago.
I personally remember being rather annoyed when our uni project in 2012 had to be built with Flash and AS3 as it was well known that flash would get the axe in a few years.
We had the option to use JS as well but it wasn't in the curriculum and 4 out my 5 person team didn't want to do any of the extra work.
Even if Flash was used, it was mostly video players and some online games (like Newgrounds). I think YouTube came out with their HTML5 player in 2012 so Flash was definitely moving into a more niche status and I don't consider it essential after Flash based sites went out of fashion (which was closer to 2008-2009. HTML5, ECMA5 and CSS3 (or just HTML, JS and CSS as it's known today) largely made flash redundant before 2012.
There;s a difference between still being around and being an "essential part" of the web.
Flash was almost certainly not an essential part of the web. 2012 was 5 years after the iPhone and by then Android and iPhone constituted the majority of web browsing, and neither supported flash at all.
You couldn't very well say something was an essential part of the web if it wasnt supported on those devices in 2012.
Loved Opera back in the day. It had "page specific settings" that would let me allow to set HTTP/Socks5 proxies on a per-page setting - today you cannot even set a proxy anymore without changing your system settings (not browser settings).
If anyone asks, I used this heavily to bypass country blocks/redirects on a per-page level and also with privoxy on some privacy invading sites.
And don't forget the ability to run a web server from within the browser. Opera was sometimes buggy, but it was really the first "browser as an operating system".
foxyproxy does this. i set up a ssh passthrough and routed foxyproxy with only specifed pages to use local proxy which meant they would be routed through the ssh. it worked for what i wanted
> It had its own, super fast, rendering engine (I forget its name, Presto?). It had a built in email client, feed reader, calendar (unfortunately with no Exchange or Gmail integration), a notes app, a powerful download manager and even a Bitorrent client. And a TON of features and UI flexibility.
And don’t forget its MDI interface [1] which made using all those features a joy and is still today better than all the tab implementations of modern browsers (for power users at least).
Just to clarify in case someone's wondering, a browser named Opera still exists and gets updated now, but neither the people behind it nor the underlying technology are the same as in the old Opera this post is referring to.
As an old-time classic Opera user since version 5 or so, I do use the current Opera because I find it somewhat better than Chrome (it has more built-in stuff, including mouse gestures, Whatsapp/Telegram support, etc. I hate the barebones browser+extensions model) but unfortunately it's miles behind the old Opera experience.
I have also tried Vivaldi, and while it's a worthy effort, I'm sure many people will love it and I recommend trying it, it's missing one of the Opera characteristics I valued the most: Opera had practically zero UI lag/latency, whereas Vivaldi is rather slow, as the parent post says.
Vivaldi is exactly as slow or fast as any Chromium-based browser. A this point I've surrendered to the notion of the web never getting faster, but just eating away any efficiency gains in computing.
I like Vivaldi for the feature set. It's rich enough for the regular power-user, but the presentation is straightforward enough that you don't feel like you need to pimp your browser like you would Emacs. I felt home with Opera, and I feel home with Vivaldi now.
> Vivaldi is exactly as slow or fast as any Chromium-based browser.
That is not my experience at all. I tried a few different versions over the years (including the beta with the mail client, but not yet this 4.0) and the UI always feels ... slow, and not quite snappy enough to be comfterable. It's really hard to describe this and I'm not the kind of person who quickly gets annoyed by this sort of thing, but for me personally it's too much and every time I tried it it was annoying enough to give up after a day or two. It's like using some remote connection with X or vnc (okay, maybe not quite that bad, but definitely reminiscent of that).
In comparison, Chromium or Firefox doesn't have this at all on the same computer.
Maybe it's a Linux thing; I don't know. I can imagine it's not the highest priority for them. A lot of it seems in the UI layer (and not the rendering layer), which is vastly different from Chromium. I hope they will improve this at some point (and I'll try out this version, too) because outside of this Vivaldi looks pretty neat.
Try the same test with Vivaldi 3.7, 3.8 or 4.0...
They've optimized the browser interactions quite dramatically so it's much faster than the video shows. But... It's still slower than Chromium's UI.
Vivaldi UI got significantly faster exactly two times from product inception. First time was a random Taiwanese hacker reverse engineering it, noticing they were resizing _all background tabs_ every time you resized browser window, and sending them a patch https://github.com/WillyYu/vivaldi_1.7. The second time Vivaldi themselves speed up Tab open operation, and only this one operation. "tabs open twice as fast. increasing speed up to 2x" sounds impressive until you measure 400-500ms instead of previous 700-900ms, versus ~16ms in Chrome (one screen refresh).
As someone who now uses Vivaldi as their daily driver I can vouch that it's a bit slow, especially on older hardware. Obviously it's not intolerable, but it would be nice to see some improvements in that area.
tldr: Tab bar is just a DIV full of SPANs, one SPAN per Tab. Pages are in a list of DIVs containing iframes. Switching a Tab is a simple process of moving "active" class from one SPAN to another and changing style of one iframe to "visibility: hidden; z-index: -1;" while another one gets "visibility: inherit; z-index: initial;". So far so ~good. So how does Vivaldi handle Tab close? You would guess closing a Tab would entail
- switching iframe visibility away from Tab being closed first - instant visual feedback for the user
- removing SPAN representing Tab we want closed
- maybe fixing widths of remaining SPANs
What actually happens is a horror show normally reserved for second semester CS students trying OOP for the first time. List of all Tabs gets traversed multiple times, most of them rewritten couple of times, everything generating individual DOM changes.
10 Tabs open and you close the first one = ~30 DOM Reflows.
100 Tabs open and you close two in the middle = ~300 DOM Reflows.
100 Tabs open and you close first 10 = ~3000 DOM Reflows.
Stuttering UI is the staple of Vivaldi, codebase is full of functions running in linear/polynomial time to the number of Tabs open/Tab position on the list, in places where O(1) is trivial. Vivaldi is a browser that used to Resize contents of every single Tab opened when you resized its window - resizing was slower the more Tabs you had in Tab bar :o, going fullscreen could take 10 seconds.
I use a browser probably 12 hours per day. Maybe there's enough people willing to pay a dollar or two a month for that to be a possible business model again.
This honestly sounds like a great idea. At this point I'd pay money for a browser that was private, fast, and dev-friendly, simply to avoid the moral dilemma of switching from FF to a Chrome clone.
FF is my daily driver (since Quantum in '17) but I've found it really just isn't holding up these days:
• Profiling functions between Chrome/FF reveals that Chrome is frequently considerably faster (usually 60%+)
• During WebGL work I'll often hit blue screens around memory management (this seems to be a relatively recent development)
• Similarly, trying to profile a WebRTC application with video feeds will consistently trigger a blue screen. (I can tell this is from FF, as the browser itself flashes white a few times before the system goes down.)
• Things like the UI changing - while minor - are enough to cause inconvenience, plus the occasional random feature removal (View Image context menu option) indicates that I'm not necessarily the target audience for the browser.
I actually slightly prefer the devtools in FF than Chrome, but there are a number of quirks that have started appearing over the past year or so, ranging from actual bugs to just weird UX.
As I write this, I'm realizing I'd love a browser that's seen more as a devtool than a catch-all web navigator which happens to have a profiler built in.
If WebGL and WebRTC are causing bluescreens, that is probably a GPU driver bug (though I realize that users don't care whose fault a crash is). If you have steps to reproduce, even if not 100% reliable, you can file a Firefox bug in Bugzilla. Mozilla has contacts at Microsoft and can pass along the bug.
Yeah, no user application should cause a blue screen of death or kernel panic under normal circumstances, unless it is either a bug or they're messing with something from the kernel (but in this last case, you generally need to run the program with high privilege anyway).
This is a distant ideal with today's GPU drivers. Browsers go to heroic lengths to rewrite graphics API calls and recompile potentially hostile code originating from WebGL to avoid tipping over crashy GPU drivers.
What normally happens with native GPU intensive apps is that devs change their code to work around driver and OS crashes, and if the app becomes popular, GPU vendors will react to frequent crashes from that app with driver fixes.
One small thing in the latest redesign:
if you have two tabs, one active, one not, the visual cues suggest exactly the opposite. The active one seems like a clickable button, the inactive one seems pushed down and not clickable.
After misclicking hundreds of times, I still couldn't train myself to go against my perception and follow the designer's "bold vision", using it feels like writing with my left hand or steering a bicycle with a crooked wheel.
This is the way. Additionally I went into about:config and set browser.proton.enabled to false, which fixed most of the issues I had with the new design.
Still a Firefox user. There is really no alternative for me that don't drive me crazy.
That said:
I'm not salty because they replaced XUL but because they mostly gutted it without replacement.
I'm salty because they pretend we are a community whenever they want donations but write as if we are enemies when we try to ask very nicely about something that is missing and has been missing for years.
I'm also salty because they really haven't been clear about donations not going to browser work but to all kinds of niche projects.
I'm salty because WebExtensions still can't do something as seemingly simple as bind global keybindings years after the transition. I actually supported the switch, but back then I was optimistic about missing features being added soon.
The new UI overhaul looks awful, they killed most useful plugins with the big update awhile back, and ideologically/politically Mozilla has become equivalent to an impoverished Google so personally speaking I'm horny for something new in the browser space
My biggest issue is, running it with a 4k resolution on Windows seems to be considerably slower than Chrome (Intel graphics). With a full-hd resolution I didn't notice the difference and on a different PC running Linux I don't have any performance issues either.
I think that came during the period of Black & White by Lionhead Studios. (which was a milestone in using gestures to cast spells by swirly your mouse around)
And Opera openly admitted that they copied B&W there, because they considered it a very useful feature. It even came in a sub-version, which at that time was somewhat unusual for such an important feature.
> right click drag left to go back, right click... swirl? to reload
I still customise my Chrome and Firefox mouse gestures add-ons to work exactly this way, despite having long forgotten where those conventions originally came from.
I was an avid Opera user around 2004. True MDI interface, mouse gestures...come to think of it, that was the peak of my web browsing.
Sure, my current browser supports the latest web standards that enable great things and sure, my internet speed is a lot higher. But just from a "browsing the web" perspective, Opera was the high point.
No, it was not. That's a popular myth, but it's wrong. There were other browser with Tabs before opera event tried them. And Operas first attempts wasn't really Tabs in the modern sense. It was just a poor working button-bar for an MDI, which was a bit cumbersome to use. They fixed it two versions later, around the time when the feature gained attention in other browsers and even Mozilla Suit got it's first Extension that became so insanly popular that it was build into Phoenix (later renamend to Firebird, then Firefox) out of the box.
Though, Opera tried many things in that area. There never really lost the MDI-Spirit and went with different approached than most other TDI-Implementations. Made it more useful for some, more complicated for others. So for certain specific functions they probablly were the first ones.
However, Opera supported an MDI interface before it supported tabs, and if you count that it may have supported multiple browser windows within one MDI window even earlier, which could be said to be a form of tabs.
Everything depends on how you define "tabs", as usual.
And in a HN thread a couple weeks ago someone posted about how “everyone” switched to Chrome because it invented tabs, if you want to know what today’s web developers think.
I loved tabs in Galeon, it's the only browser I remembering having it at the time, but I probably didn't try Opera. I was sad when Galeon went away (or removed tabs? I don't recall) and I couldn't find another browser with tab support for a while. Again, I may have overlooked Opera :)
Galeon was also the only usable Linux browser having a native look on Gnome (Epiphany was the other one, but I remember it having more difficulty with a lot of sites).
It embedded the Mozilla engine (which was a supported use case back then, before Firefox) but launched a lot faster. Usable on a 200Mhz machine at least.
Opera had some ridiculously powerful keyboard shortcuts and performance as well for the time. I used to play this web game years ago and past beginner level the only way to compete was to use Opera because you could use a combination of keyboard shortcuts to send attacks/defences from multiple tabs within milliseconds.
> Opera was first a paid browser (which limited its reach)
I'd argue that this, in combination with easy availability of pirated serials (and the actual quality of the engine, of course), contributed to Opera popularity because "why use freeware when you can use premium software (which other people supposedly even pay for) for free".
I believe this works so if I were going to release a commercial app I would make sure to put some serials on pirate sites.
Even today, if you release a new Chromium-based browser, make it paid but easily piratable it probably is going to enjoy more popularity than if you release it for free.
Not to forget Opera Mini which did server-side rendering for mobiles. Results were WAP-like quality and mostly sucked (by nature of device's capabilities back then).
I disagree. Opera mini allowed me to browse the web almost normally (I was using a Nokia E65) and I never really understood why I regularly see people saying the iPhone brought internet to smartphones or something of the sort.
I also used Opera on my old Symbian phone. Probably the number of iPhones after two or three years was larger than the number of mobile devices that people actually used to browse the web before the iPhone. Nevertheless where I live (Italy) the number of Android devices were always greater than the number of iPhones. I'm thinking about the Galaxy S and S2 in the very early 2010s and the other flagship phones of the day. So, the iPhone brought the touchscreen to the masses (the iPod too), Android brought the mobile internet.
Yes I give them touch screen. I admit I was a touch screen skeptic (the kind of person who became mad when someone would touch a screen so doing that on a regular basis seemed a really bad idea to me ^^) and did not get a phone with a touch screen until much later, when they were basically the only option :)
Disagree. In 2007 I used a Nokia 6300 (https://i.imgur.com/5ezXN1R.png) and I had an unlimited data plan for €10/month (because, who was going to use a lot of data on 2G on a small feature phone?).
I used to smoothly browse the web on it using Opera Mini. It had a simulated mouse and rendered most stuff excellently.
Also, during the days of Symbian OS Opera Mini was the best mobile browser available. I remember logging into my bank account and actually doing a transaction on my Nokia N73 running Opera Mini.
Not my experience. I use it out of nostalgia, but only for select sites. Many sites with modern SSL config won't load. Click a random link that lands you on a js-heavy site and the browser will stutter or often crash. Which is impressive because back in the day opera once had the fastest js engine by a mile. Goes to show how far V8 has come.
But browsing traditional sites that just have a little bit of js and are mostly static still feel faster to navigate on this 8 year old browser than the most recent chrome.
The Vivaldi mail thing must have been in the works for a while. Since very early on, they offered an email address at their .net domain. It wasn't integrated or an email client, but they had that aspect setup early on.
They licensed the browser for devices like set-top boxes and similar devices. There was also money coming from Google (for using it as a default search engine).
I guess they had also some business model built around Opera Mini and relations with mobile providers.
Vivaldi is developed by a team created by the former Opera founder and CEO. Vivaldi kind of tries to re-create the spirit of Opera, but I think it's going to be too hard to do it.
For those who didn't use it, Opera was first a paid browser (which limited its reach), then an ad-supported browser (which again limited its reach). It then finally became a free browser but by then it was too late.
It had its own, super fast, rendering engine (I forget its name, Presto?). It had a built in email client, feed reader, calendar (unfortunately with no Exchange or Gmail integration), a notes app, a powerful download manager and even a Bitorrent client. And a TON of features and UI flexibility.
It was super compact, a marvel of engineering and UX design that managed to pack all those things in a package of about 5MB at the time, and you wouldn't even see or load the extra functionality like the email client if you didn't use it.
Unfortunately with HTML5 and the Chrome-ification of the web, it couldn't keep up :-(
Vivaldi tries to do the same on top of web techs and web techs just can't handle it. Web techs are almost as flexible but they're really slow and bulky.
Still, I wish them luck.