Posscon in review

Attending Posscon was definitely worthwhile.
It was exciting seeing the RepRap 3D printer in action since I've followed news stories about it for a long time. Neil Underwood revealed a lot of things that hadn't been in those stories. For instance, although the RepRap team wants to make a printer that can copy itself, they have realised that the magnets in the motor make that difficult.
John Diamond was very knowledgeable about Computer Aided Design programs. He also works on open source video games. At the moment, they are technologically pretty far behind mainstream titles. This is mostly because open source game projects tend to have only a few contributors, and those do it as a hobby. They are catching up, though. The primary advantage of open source game development is that knowledge is shared between projects. They also tend to develop only for the PC, which takes the complications and limitations of consoles out of the equation. I think that real time ray tracing will be a boon for these development groups, since the software involved is actually much simpler than current technology.
I also talked with a couple of people who were interested in hiring programmers, which was cool.

POSSCON

I'm looking forward to hearing talks from some of the biggest figures in the field of software development. Three presenters that I am interested in meeting are Jim Jagielski, John Diamond, and Steven Jackson.
Jagielski has a lot of experience in open source programming and cloud computing, which are both only growing in importance. Diamond has worked with CAD applications and open source physics engines, which have always been interests of mine. Jackson's work with genetic algorithms and data prediction also seems interesting.

Banshee continued

Our first submission to the Banshee project was largely successful. The main developers wanted several locations in the code that referred to "banshee-1" to be changed to "banshee" without breaking anything. With a bit of help, we decided how to handle each instance and submitted a patch. The patch has been reviewed and incorporated into Banshee.
We are now working on an extension using Gwibber that will allow people to post statuses to Facebook and other social networks from within Banshee. By default, these statuses will contain metadata for a selected song. In addition, the user will be able to edit the message before it is posted.
I know a lot of people who use both Banshee and Facebook. It is interesting to think that I might see this extension we are writing in use.
For the last few days we have been struggling to set up the project, but the issue has been found: In some builds of Banshee, autogen.sh was still looking for "banshee-1". Apparently the switchover caused a few problems. This has clearly demonstrated the point that even small changes can be hard in a large project. It seems challenging to balance a "design first and code second" philosophy with the rapid iterative development that agile processes call for. I imagine that most large projects require difficult revisions from time to time.