All day long I’ve had my hands full working on a personal project, I’ve been beating my brains out with an algorithm involving a couple of complicated calculations, in fact, I just got it done. The code (written in C#) seems to work good enough, I’ll have to clean it up a bit, though. There are still quite a few useless source lines of code. As I need to do some C# + OpenGL programming for my project, I’ve been trying out some available OpenGL bindings for C# like opengl-sharp that includes a Gtk# widget and GtkGL# that includes it too. Finally, I decided to use Tao Framework which is a collection of bindings to various native libraries (OpenGL, SDL, OpenAL among others), furthermore, it seems to be the current standard OpenGL wrapper API for Mono.

As Tao is not a Unix technology but a migration technology (from .NET), perhaps a whole bunch of people may think it’s not a good choice for building a GNOME application. Nevertheless, I, myself, think we’re in the real world and Tao is, for the moment at least, the best choice for developing cross-platform OpenGL applications.As part of my project, I’ve been using the F-spot’s source code, so I got a modified F-spot version (0.2.0) running on my laptop. My F-spot-0.2.2 (unmodified version) runs perfectly, but when I modify it with the same changes I made to F-spot-0.2.0, it simply crashes. I need F-Spot-0.2.2 to work with the new features I added to F-spot-0.2.0 (3D modeling from pictures) in order to continue my project so I’ll have to patch it once again. It seems that this may be due to a dbus-sharp problem (Managed D-Bus issue) which is the C# implementation of D-Bus used by F-spot-0.2.2 for single-instance detection. I’ve not had enough time to read the D-Bus API documentation so I haven’t the faintest idea how those lines of code work (within F-Spot-0.2.2). Anyways, the algorithm I’m working with for generating 3D human faces from pictures doesn’t work yet due to the lack of important features within the skin chrominance model that I use for creating realistic 3D human face models, in some cases it works but it takes up to 10 min of computing time ! so I still need to fix a bunch of issues.

In the future, I’d like to add some extra features like statistical modelling and host the project in a subversion repository. I started hacking on this project as an educational exercise, now I’m really getting so excited about it.

By the way, I wasn’t going to post a geek-oriented blog entry, why did I do it?