Google Summer of Code 2011: what’s trending?

Google Summer of Code 2011: what’s trending?

Last night Google announced projects that were accepted for Google Summer of Code 2011 program. The amount of graphics and multimedia related ones is rather large, you can see them all here. Instead, let's talk about trends.

Inkscape has relatively smaller amount of projects, just four this year, and three of them deal with long time annoyances: far from perfect use of CSS, rendering performance (second stage of last year's project by same student) and lack of good UI for handling JavaScripts. The organization has seen its share of "let's do some entirely new things" in the past years, and it's the second year where Inkscape is working on mostly fixing things and refactoring code. The fourth project, KML support, is going to close another gap — more digital mapping related features. Two of the four students have already successfully participated in GSoC last year.

Scribus goes to GSoC with just two projects — further work on new Scripter (which is not so critical) and basic tables implementation which is really long overdue. Hence, Inkscape and Scribus take a modest amount of projects and, essentially, tie loose ends.

GIMP takes a considerably different approach and goes for both fixing things and innovation. Only one project, GimpSizeEntry rewrite, is completely about refactoring which you would expect to be a focus given the plans to finally do high bit depth in v3.0.

Two more projects, both by a former GSoC student and another person who already contributed to the project, are to introduce new tools. Adaptive image cloning will make it easy to seamlessly place an object from a photo to a different photo with different lighting conditions (white balance, exposure etc.). And iWarp filter as a tool is something everyone's been waiting for a really long time.

The last two projects assist transition of GIMP to GEGL by porting filters to GEGL operations and bring GEGL to state of the art rendering by introducing OpenCL support.

Krita team (part of KDE org) this year is focusing on features that, subjectively, will bring it in line with current GIMP: SIOX based foreground selection tool, PSD importing and exporting, tagging and management of resources (brushes, gradients etc). For PSD support Krita is likely to surpass GIMP, because it simply has more features to map Photoshop features to. Another project will add threadsafe tools execution. The overall focus is interesting given "rebranding" the project for digital painting last year.

Blender, with its whopping 17 projects, attacks several targets at once. It's very easy to notice that three of the projects deal with camera tracking which is interesting given plans to use the technology for the next open movie project, codename Mango.

The other targets of attack are paint system (three projects), animation (two projects), game engine (three projects) and automatic UV seaming (two projects). Other projects deal with diverse topics: retopology, 3D audio, i18n.

Digital photography this GSoC is a topic of several more projects:

  • darktable goes to GsoC with one project to fix various UI issues;
  • digiKam is going to feature simple panorama stitching and a clone tool;
  • hugin is going to get straight lines drawing tool, and Enblend will feature optimized seam finding;
  • Shotwell will get improved slideshow and positional tags (tagging people and face recognition).

There are even more interesting graphics related organizations participating: Point Cloud Library and OpenCV, both with 11 projects, as well as OpenImageIO library with 3 projects. OSGeo is participating with a little over 20 projects too.

It's interesting to see PiTiVi participating with three projects:

  • automated compositing in PiTiVi using optical flow with GStreamer;
  • implementing render profiles, video uploading and GUI enhancements;
  • animated effects, clip transformation and titles.

Another GStreamer project isn't, strictly speaking, a PiTiVi project, but "plugins for performing automatic video correction, such as noise reduction, colour correction, or camera shake correction" sounds very much related.

PiTiVi is the only non-linear video editor for Linux that participates this year (one would expect Kdenlive joining KDE for that), and the projects intend to both implement basic things and lay foundation for more sophisticated features.

Very few audio organizations are participating this year. Both Audacity and Ardour decided not to go for GSoC this year, since both expect a new stable release later this year which is somehow more important. So we are left with an attempt to give Jokosher a kiss of life, four XMMS2 projects and three Mixxx projects — all dealing with improving what's already there and adding long overdue features. It's also worth noting that KDE organizations takes this year two students to work on Simon — a free/libre speech recognition engine.

All in all, it looks like organizations very well understand what users expect from them and go for projects that bring respective applications up to date with modern technologies and workflows.

Was it useful? There's more:

0 Responses. Comments closed for this entry.