Starting small...

I've set up the SVN today, turned out to be very easy. Setup was straightforward, I just needed to add a few lines to my Apache2 config, and I was already able (via TortoiseSVN) to create a basic repository. I cleaned up the old development dir (it was full of temorary build files and such stuff) and checked in a clean installation into my local SVN. Seems to work fine. First operation was a tag, 0.0.1, so I have the starting point fixed. I'll make the GLExtensionHandler class public soon, it contains the full OpenGL Extension registry (as bundled with GLEW), and a Python script (which requires wxPythonfor the GUI, but nothing else) that allows you to do two things:

  • Create a XML file (documentation of the format is available somewhere, will dig it out) which contains a structured extension registry
  • Create a C++ extension loader class. This is basically a large init class, which takes care of checking whether a particular extension is supported, and if so, loads all the functions associated with the extension. It also adds defines and typedefs when needed - all you need to use OpenGL extensions

It has been tested already under Win32, and it worked, so I'm pretty confident that some people will find it useful. My Boost Installation also needed an update, so I got the 1.32-2 self extracting exe from, and after compiling for my build target (vc-7_1) I got exactly 592.438.460 bytes of libraries - all without a single error. Great work, guys! I also wanted to install Trac, which looked like the best solution for bug tracking, documentation and such stuff for me, unfortunately, the web page seems to be down.

A quick review of the current codebase

Before I start with any kind of schedule whatever, I wanted to take a look at the current status of the code. Most of the math library is checked in and pretty well tested (covered by some unit tests even). The filesystem - although rough - is also in place, and mostly documented. However, the whole scene graph etc. is in a largely untested, unstable and unfinished state. This is clear, as this was by far the biggest part of the source, and we tried to get the engine running as soon as possible, without too much code that would be not needed at the beginning. Anyway, this area will need my particular attention before I can start with the design process again.

Decision ...

So, here it is, my decision: I'll dig out the 3D engine again. I think this is gonna be the best solution for me, quite interesting, and as I'm getting a new rig, I think I'll be able to test the more interesting stuff that has mostly not been implented yet. My current plan looks like this:

  • Set up a local SVN server
  • Understand the Aqsis MP Dump format. I've written a quick reference today (check it out in the Wiki at The MP dumps are cool, as they contain usually huge (millions or more polygons) which saves me from creating complex test scenes on my own - at least for the beginning.

After that, and when my new rig finally arrives, I'll start coding again seriously. So much for now.

Some random thoughts

This is the first "real" post, that is the first one that is not meant to test the blog :-). At the moment, I see two choices, both intersting, but I can't really say what I will choose today, will have to sleep a night over it. The two alternatives are:

  • Start off my own company and write an innovative web application. While this choice has some commercial potential in it, I tend to believe the marketing and such might be the main problem, plus it is doubtful I'll be able to continue work (especially support) on it after I start my studies again.
  • Resurrect my 3D engine and start again with realtime 3D stuff. This is what my heart tells me to do... No profit in sight, though.

One way will give me money and hard work, the other will give me hard work and satisfaction. As I said, some thinking will be necessary before I can decide - and as this decision will have long term consequences, it has to be done with special care.

New logo

Finally (:-) 10 minutes later) I found a suitable picture. It's a detail of a photo I've been taking today in the morning (nearly missed the bus to university because of it). It shows our local river in some extreme ground fog.