KDE Blogs 

Google Summer of Code application deadline is just a bit more than 24 hours away and we are still looking for highly motivated students to work on a Marble Virtual Globe project this summer. Tomorrow is a holiday in many countries, so you might still have some time for the application. Make sure you file your application not later than tomorrow at 19:00 UTC. The last project topic for GSoC that I'll cover is "OpenGL support for Marble": You might ask: Why OpenGL? Doesn't Marble make use of OpenGL already? No it doesn't. Marble currently uses it's own software rendering to provide the different projections. Both texture and vector data are fully rendered in software by default. This works reasonably fast on most devices, but compromises must be made with respect to the map quality. OpenGL, however, offers higher-quality results, is more state-of-the-art and perhaps helps to reduce power consumption on mobile devices. It therefore makes sense to introduce an OpenGL mode in addition to the sotware rendering mode in Marble (the user should be able to choose between software rendering or OpenGL). The main task is to refactor the Marble codebase such that both the software rendering and the future OpenGL code share as much code as possible. In particular, visibility control (which objects should be rendered according to the current view parameters) should be factored out into separate classes, such that they can be reused in the OpenGL mode. Initially the primary focus will be about replicating the current feature set and behavior of Marble. So the globe would still be browsed in looking top-down. Later on one could extend Marble and introduce bird-view, camera flights and "real" mountains and "real" 3D buildings. Bernhard Beschow created an initial prototype of the OpenGL mode in an experimental branch already. You can look up more details about this prototype in his blog. Here is the video and a screenshot: Marble - OpenGL Prototype Since then Bernhard has partially prepared the current Marble master branch for inclusion of OpenGL. But there are still lots of missing bits and pieces in order to have the actual introduction of OpenGL inside Marble. Expected Results for this project: duty: refactored code that separates rendering from visibility management "icing": an OpenGL mode If you are interested in this wonderful project then you should act quickly: Deadline for applications is on Friday, April 6th, 2012m 19:00 UTC. Apart from the usual GSoC student application guidelines your application should: describe the benefit of the feature from a user's perspective (including self-created mock-ups and screenshots) provide a rough technical explanation in your own words what the project will be about. state why you are the best person to master this project. So if you are a student then we are looking forward to your application! Don't hesitate to ask us any questions.

Google Summer of Code application deadline is near and we are still looking for highly motivated students to work on a Marble Virtual Globe project this summer. This is a follow-up of yesterday's blog about OpenStreetMap vector rendering with tiling support. Today I'd like to describe another important feature that we'd like to see covered as a GSoC project. In terms of technology this project is very similar to the "OpenStreetMap vector rendering with tiling support" topic. In fact if you applied for that project already then you might want to apply for this one as well if you want to increase your chances: The topographic "Atlas" map is the oldest map theme featured by our Marble Virtual Globe. The original aim behind the Atlas map was to create a quite detailed map from very little data. The data would get shipped together with the Marble application for offline usage. Read More

Lots of KDE SC 4.8.2 packaging and testing today. But some folks can't test it easily because they have small bandwidth or they don't want to install experimental packages on their local computer. I've been using EC2 machines for a while, which you hire by the hour from Amazon Web Services. Whenever someone in Kubuntu needs a machine to compile or test and install on I run my script to get one going and give them access. They're also handy for showing people how to do e.g. packaging, you can log into the same machine and share a screen session to see what the other person is doing. Read More

Google Summer of Code application deadline is near and we are still looking for highly motivated students to work on a Marble Virtual Globe project this summer. One of our favorite topics is "OpenStreetMap vector rendering with tiling support". And in this blog I'd like to describe our current ideas about this interesting subject a bit more in detail: Basically all the maps displayed by Marble are based on bitmap texture data. Some of these maps even consist of multiple texture layers. On top of these texture layers we display further content like placemarks and possibly even simple geometries (like GPX or KML data). For displaying this bitmap texture data efficiently we are making use of the popular concept of QuadTiles: the data is arranged in different zoom levels. Read More

And Now for Something Completely Different. Kexi's Easter Egg just appeared and quite magically it's only available on April Fools' Day ;) To try, run any beta of Kexi 2.4 or 2.5 alpha. Then close the app, run again and look closer. Enjoy! Read More

As I wrote in my previous blog post, I've been doing work on getting Calligra working on Android. My first attempts of this used Calligra Mobile, our QWidget based mobile UI originally written for the Nokia N900. One big problem with this is that the scrolling speed when running on android was horribly slow. I think I got maybe 1 or 2 frames per second, and that on fairly powerful hardware. Early on in the development process of the calligra based document viewer for the Nokia N9 we were having similar performance problems, caused by wrapping our QWidget based canvasses in a QGraphicsProxyWidget. Now Calligra Mobile itself is completely QWidget based, so it shouldn't have those problems (it works decently on the N900 afterall), but because of the way Qt on Android is implemented there is a similar kind of indirection going on as with QGraphicsProxyWidget, so instead of trying to fix the performance of Calligra Mobile, I decided to instead give one of the other mobile UI's of Calligra a try. Read More

Since so many students have already shown interest in working for KDE during this years Google Summer of Code, with some already having started to send in first drafts of their proposals, I thought I'd a bit of advertising for a specific idea. Well "my" idea :) Read More

Since I had to announce that Canonical was dropping support for Kubuntu from 12.11 (and then had to announce two days later they were dropping support for 12.04) I've been getting lots of people asking "is this the end of KDE?" Read More

I've been playing with getting Kubuntu Active on ARM. Getting a working ARM setup is a lot like getting a working Linux desktop setup when I started in 1999. It's unclear what computer you need, it's unclear what install image you need, it's unclear how you install it and then it doesn't work and it's unclear how you debug it. For some unknown reason Ubuntu Desktop images from precise don't work on my Pandaboard but from oneiric they do. Ubuntu Server from precise seems to work so I've installed that and installed the Kubuntu Active packages on top of it. Maybe soon we'll have working Kubuntu Active images on ARM. Read More

The first Ubuntu flavour for tablets is now making daily builds. We even got our first bug reports from our localy Plasma Active upstreams. Images are for i386 only for now, ARMv7 should be added when we know it's a bit more stable and have testers. Read More