The New KDE "Fiber" Web Browser Deciding Between Qt WebEngine & Chromium

[phoronix]

Phoronix: The New KDE "Fiber" Web Browser Deciding Between Qt WebEngine & Chromium

A few weeks back we wrote about Fiber, yet another web-browser for Qt/KDE, while today there's a bit more information...

http://www.phoronix.com/scan.php?page=news_item&px=Fiber-Web-Browser-Engines

[shmerl]

They should go with Servo.

[Azrael5]

I hope for QtWebengine it would be the first, besides matching KDE.

[Maxjen]

Definitely CEF, because Servo.

[magika]

Is the difference between CEF and QtWebengine that they will have to re-implement all Blink integration parts that Qt already did?

[schmidtbag]

As a KDE user, I don't care - I won't be using it anyway. I found Konqueror to unnecessary clutter. KDE really needs to stop pretending to be a Windows clone (where Konqueror was like IE and Fiber is probably supposed to be like Edge). There is so much KDE software that could use polishing.

[Daktyl198]

I'd say to go with the CEF, so that later down the line people can build it with either Chromium or Servo, based on their preferences. It would also be a great test bed for Servo's CEF support when they get around to implementing it, as I don't really know of any other FOSS browsers using CEF.

[zanny]

As a KDE user, I don't care - I won't be using it anyway. I found Konqueror to unnecessary clutter. KDE really needs to stop pretending to be a Windows clone (where Konqueror was like IE and Fiber is probably supposed to be like Edge). There is so much KDE software that could use polishing.

While I agree that given the status of Rekonq it seems obvious KDE does not have the resources to maintain a first class browser project themselves, this also is not some ubiquitous movement from the KDE community towards using Fiber everywhere. Its one guys project, and all the involvement KDE has is that its ready to sanction Fiber as its kde applications browser if it matures.

Its still up to everyone if they want to contribute to it or use it.

Fundamentally the problem is more that we don't have a good Qt based web browser that integrates well in KDE. Well, we do, in Qupzilla, but it is also not a feature complete browser that can actually replace Chromium or Firefox - it is still missing html5 video with mse, it doesn't support webrtc, and its lacking a webapps framework anything like the Firefox or Chrome app stores or addon infrastructure. And even then, Qupzilla is using Qt and reinventing frameworks rather than using frameworks itself, mostly because its much older than frameworks and was understandably unable to just pull in kdelibs when it was a monolith.

There is also a bit of a philisophical angle here. Debian recently showed off how Chromium builds were downloading proprietary Google code to do Google Now integration - effectively capturing your microphone when the browser is open and sending it to Google, and Firefox has integrated pocket which is a proprietary 3rd party service directly into their browser. Both camps are violating my software freedoms by pushing proprietary software and its impossible to be complacent in using either of them today. So of course there are people interested in trying to fix that, especially in the KDE ecosystem.

What I don't get is why Fiber is a thing at all. Why not just fix Rekonq? It had adblock, it had syncing, it had bookmarks and password storage and keywords and when it was being actively developed at its peak in 2012 it was very competitive with Firefox and Chrome of the time. I'm not sure if Fiber is using QML for its interface, but if it is not Rekonq had a lot of good code that is now rotting. I'd understand a rewrite if you are redoing the UI in QML and using a different webengine, though (Rekonq used qtwebkit).

[Kver]

Is the difference between CEF and QtWebengine that they will have to re-implement all Blink integration parts that Qt already did?

A better way to look at it is to see them both as abstraction layers on top of the fast-moving Chromium engine; WebEngine is simply a Qt-oriented version of what CEF does a little more generically;

CEF is more complex and (of course) doesn't have Qt goodies like signals and slots built-in, so it's going to cost extra time to add them where things need to connect up in the browser. On the other hand it offers more integration options, may be more efficient in areas, and of course it opens up Servo as a renderer option. Since Servo is going to use CEF as it's primary API (unlike Chromium which has CEF as another layer) this means Servo might also see a much more efficient CEF implementation.

[Azrael5]

I know there are 3 main web graphic engines: geko (firefox) blink/webkit (chromium/chrome/ and many derivates) and Qtwebengine. KDE is based on Qt there are no browser engine currently written in Qt web-engine so would be the best improvements to realize the first browser by this webengine an excellent alternative to geko and webkit making Linux browser FREE... further benefit is for linux operating system based on KDE getting a good compliant web-browser.