300 million users strong, Opera moves to WebKit | Hacker News Reader

300 million users strong, Opera moves to WebKit

my.opera.com

181 pointsandreastt50 comments

As someone who's been a longtime Opera user before switching to Chrome (buggy release around 10.x, weakening site compatibility) I'm so glad to read this. This decision shows the quality of engineers who work at Opera - not afraid to throw a huge chunk of code away if it doesn't serve the company anymore the way it was intended. Thinking about it I can't make up a single example of a large software house doing such a thing.

Looking forward to get my neat features again, such as fast forward, hotkey bindings, Opera turbo..

Same experience, left Opera for Chrome for exactly the same reasons. However I miss much of the features in my daily use, the ones you listed and many more. Especially Opera's tab preview that doesn't seem to have an equivalent in Chrome is something I have found myself missing deeply. I really hope Opera becomes a viable alternative for me again now.

Back in the day, Netscape did a full rewrite between major versions. There's a Spolsky blog post about it: http://www.joelonsoftware.com/articles/fog0000000069.html

Very interesting read.

This is totally different. They're not re-writing their application from scratch, but replacing a core component with a popular, well tested and widely support one.

Indeed it is -- GP only mentioned he didn't know any example of a large amount of code being thrown out, so I provided one.

They are ditching their UI toolkit and dev tools as well. It does look like a complete rewrite, maybe with copy&paste of some old features like gestures.

Thanks for the link, it's indeed a good read.

The web may not be fully open, but it is far more open than the closed world of "apps".

(Of topic, perhaps...)

I hear this a lot, and I think, for the most part, web applications aren't remotely open (in general). While I can use any computer to connect to a web site, very few sites actually allow me to get my data out of them. In short, in my opinion, open access to data trumps open access by various clients.

On the one hand, we have systems that I can connect to with any OS I want but where my data is completely out of my control. On the other hand, we have walled gardens where the data sits, literally, in the palm of my hand. Both situations are ugly, but in the former, if I want my data, I have to pray the developers left me a method to do so. In the latter, it may not be trivial, but it is almost always possible[1].

I wish we had truly open systems across the board. Until we do, I vote we stop calling web systems open, because they are only open in terms of access, and that's not good enough.

1. For example, jailbreak, log in, and scp a SQLite db. Certainly not trivial but at least doable. In no way would I actually call this "open". I just think it is less closed than most web applications.

You still have "to pray the developers left" you a method to get the data in either case as native apps are simply the client-side of an application (just like a web apps).

The "openness" of the web is mostly referring to the technology stack (open source rendering engines), the standards, and the freedom of distribution. All of those points are not fully "open" and that's what I think they're referring to when they say that "the web may not be fully open".

Terrible news.

This means basically giving up on standards-based compatibility.

Not bad news for Opera - maybe good news for the browser - but an ominous sign on what the web is becoming.

Opera was a minor force in the ecosystem. The day to start worrying is if/when Mozilla switches to WebKit. (I prefer WebKit browsers personally, but I think monoculture is a bad thing long-term.)

Whatever Opera do doesnt cause us to give up anything.

No, but it encourages other developers to do so, which will encourage other browser makers to do what Opera did. This is a massive step back for real standards as protocols, toward the bad old days of fake "standards" as implementations.

Interesting. Opera makes most of their revenue through licensing their rendering engine. That's how they've turned their ~2% marketshare into $180MM in revenue (compared to, say, Mozilla's ~26% and $300MM).

I wonder how this will affect that.

[EDIT: I mixed up currencies; fixed it.]

A lot of the work that goes into adapting a browser for various chipsets and devices (TVs, STBs, etc.) is highly configuration dependent: Almost no device is the same. There is still a market for customizing and making device specific browsers even after Presto.

I find those numbers hard to believe. Are you sure your currency conversions are correct?

Likely, Opera Software had a revenue of NOK 900M in 2011, and USD 1 ~ NOK 5 yielding a ~USD 180M revenue. So his comment is only correct if you compare without conversion (Opera Software is indeed a billion a year in revenue, but that's a billion NOK)

I'm not sure if this will change much for end users. I'm a long time Opera user and some sites will just lock you out unless you use Chrome or FF even if Opera likely has an implementation that is more standards compliant than both browsers.

Even sites like Udacity's course viewer did this (maybe they changed it, I haven't been there in a week).

So while Opera might be using the same rendering engine as Chrome, you'll still get locked out because web apps are setup to investigate user agents to determine who gets in.

I'm not sure if I like or dislike this change. I feel like testing sites between Opera, Chrome, FF and IE I had a better chance of eliminating all rendering bugs.

Opera I feel was/is the best candidate for ensuring your code was correct.

Yes, website compatibility is one of the major issues of Opera. The second thing is lack of support in terms of add-ons and such, but I don't see how they can help that. This might just get more people to use Opera as their desktop browser though. Every time I test a website on Opera, I'm like "I should really give this browser a chance". But every time that I try it for an evening, I notice the shortcomings in support (by websites, add-ons, etc.).

Back then I was used to having fifteen add-ons in Firefox, but after Mozilla made one bad choice after another, I moved to Chrome and live by Adblock Plus and some own userscripts and bookmarklets now. Perhaps Opera isn't as much of a culture shock anymore to use, especially with the Webkit engine.

I do wonder if they considered Mozilla's Gecko engine. It has a smaller userbase I think, but does that automatically mean it wouldn't be the right choice? They don't say a word about this.

All in all, I do applaud the change. It's a hard step to take after you've been working on your own engine for years, but they've done it, and it will probably add a lot in terms of website support.

> I do wonder if they considered Mozilla's Gecko engine.

Gecko has a lot of legacy, and to be fair, even Mozilla has tentative plans to transition off of gecko, to be replaced by servo: https://github.com/mozilla/servo

I don't think it has to do with servo, which is really just a research project.

Gecko is a lot harder to embed (so yes, legacy is and issue here), and webkit is compatible with more sites (as sites are now often "designed for webkit" and use specific webkit-only extensions)

Makes the choice very easy and logical IMO. Not really good for the web, but maybe not a bad idea for Opera. Time will tell.

> Gecko is a lot harder to embed (so yes, legacy is and issue here)

Gecko embedding is also unsupported since May 2011: https://groups.google.com/forum/#!topic/mozilla.dev.embeddin...

I always liked them prioritizing to give you a complete out of the box features. Firefox's plugin hell made me crawl up the wall every time I saw the dreaded 'you must now restart Firefox' dialog. I assume it's better now but a few years ago Opera was so much ahead of FF in usability it wasn't even funny.

Many FF addons are now restartless.

We can hope that someday IE will make the move too.

Given the way MS is embracing OpenSource ( git, TypeScript ) that day may not be too far.

I guess this makes sense if you look at it from a phone/tablet manufacturers perspective: A manufacturer wouldn't necessarily have the desire to implement their own browser on top of webkit so they want a company (rules out Mozilla) that will support a web browser on their phone without also being a competitor.

The problem I think is that the non-aligned phone manufacturers are going the way of the dodo. You have Nokia committing to WM8, Apple on iOS with Safari and most others committing to Android. In theory a manufacturer could go Android + Opera but why license Opera when you can get Chrome for free?

That being said: Opera has always done a great job fighting against massive odds. I hope this move gives them the breathing room they need to keep going.

More good discussion at this submission: http://news.ycombinator.com/item?id=5211953

A wise move for Opera. Now they can hire webkit-engineers from anywere without heavy training on their model.

This is a good thing for everyone that uses Webkit browsers.

what exactly defines a "monthly user" for Opera and how is that measured? Is it based on usage statistics, or is it just number of installed/downloaded clients?

Usage. Installed/downloaded is very high.

The number says 300 million, but my actual website stats, of over 1 million uniques per month, show it less than .75%

That is why it is dangerous to draw conclusions from small sample sizes, and why useful samples need to be random. Your website does not attract random people, so it is not an accurate sample. Opera usage varies massively by country for example, is your website available in Czech? If not, your sample will self select to have fewer opera users.

Hmm, any bets on how long before the folks at Mozilla make the move as well? It seems inevitable at this point.

I don't think Mozilla will ever switch. The reason why they've their own engine and try to keep a decent marketshare (which they do indeed have), is to be able to have a say on web standards, and the web in general.

Having a say ensure there isn't a single voice taking decisions that led us to things like IE6. And we're actually starting to see IE6-ish stuff happening with webkit, as it's nearly the only rendering engine for mobile.

One vendor != good for standards.

Oh also, their engine happens to be pretty close to webkit performance wise.

I would think one vendor would be great for standards: it is the standard. It sucks for any kind of innovation, though, as IE6 shows incredibly well.

Out if curiosity, what IE6-ish things do you see WebKit doing?

> I would think one vendor would be great for standards: it is the standard.

Well, IE6 was the standard. See how many sites were "optimized" for IE6 and getting them to render in a browser that supported these pesky underdog W3C standards was bound to fail.

> Out if curiosity, what IE6-ish things do you see WebKit doing?

Not quite the same, but the -webkit CSS prefix that people use without a fallback for other browsers: webkit browsers display the stuff and on e.g. Firefox or Opera it looks broken, because nobody cares to add a -moz or -o prefixed version, or even the non-prefixed version.

It's not one vendor it's just one (open source) implementation. I'd argue this is better for innovation as new features will be quickly ported between the vendor forks and spend more time improving the product than catching up to other implementations.

> Out if curiosity, what IE6-ish things do you see WebKit doing?

You mean like shipping various non-standard features and advertising them to web developers? Not like other engine's aren't doing similar things in various cases (though WebKit on iOS is shipping features Apple is actively opposing ever being standardized, backed by patent threats, which is a bit different from most browser vendors).

But the real issue is not what WebKit is doing, it's what web developers are doing. And the IE6-ish thing web developers are doing with WebKit is creating sites that deliberately only work in WebKit.

It isn't like WebKit is intentionally doing IE6-ish things, it's just that any rendering engine will have certain bugs. Developers try to work around these bugs, and if their is only "one" implementation, they'll start relying on its quirks. A good example in C would be storing a pointer in an int. It made perfect sense at the time, and it allowed people to be lazy. However it doesn't make any sense these days.

I believe they've said never. They don't want just one major rendering engine.

Doubt it. Unlike Opera, Mozilla is a non-profit. There're not shareholders pushing for higher profits, regardless of consequences.

This will never happen.

What will be the first release with Webkit as the rendering engine?

My guess would be the next major release, that is 13.00 for Desktop and Mobile.

I feel compelled to jump in and give a warning about Opera. I also know that the herd mind will probably pound my comment into oblivion, which is why I made a new account.

Opera is scammy, unreliable software. DO NOT DO MISSION CRITICAL WORK WITH THIS BROWSER. It will flake out on you in so many uncanny ways its not funny.

Also, Opera had the weird habit of becoming my default browser on my Win 7 machine repeatedly, despite me never doing so and only referencing it for testing purposes.

That, along with Opera having the top 3 spots on HN's front page further indicates to me that Opera is not for real. Like I mean, come on. How many of these submissions and upvotes are fake.

Opera, you're not on the up and up and you know it.

TL;DR: Don't trust Opera.

The reason why you made a new account seems to be because you're not giving any arguments for your warning. Opinions with reasons are usually upvoted by HN.

I personally never had Opera become the default browser, nor did I have any other problems than compatibility with it (which is normal, given their market share). I also don't think the upvotes are fake.

So, why exactly "SHOULD WE NOT DO MISSION CRITICAL WORK WITH THIS BROWSER"?

Perhaps because it's all caps? I can see no other justification :)

1. No arguments, just option.

2. Herd-mind might apply for Chrome-Hipsters, but Opera is far from being mainstream. Sadly.

3. I'd love to have Opera as my default browser everywhere, but it just doesn't work that easily for me. Maybe you can share me your weird/unusual insights on how to apply a default browser for the entires win7 system, including all applications?

4. You must be doing something horribly wrong. I DO MISSION CRITICAL WORK WITH OPERA - all day long - for many many years.

5. Shame on you for creating a fake account. You deserve to be downvoted.

6. tl;dr is supposed to be _above_ the content.

Reply on news.ycombinator.com