Haven’t said much the past year. I have been moving to other projects that don’t have as much blogging interaction.

A year ago, I did a bunch of playing with robotics and ROS2. This culminated in a small, multi-node setup with a camera that tracks faces. What I learned was captured in a presentation that I gave at our local ROS Meetup.

After that, I played with the Basil virtual world viewer by creating several OpenSimulator modules that convert and present a region to Basil. This included Loden which converts OpenSimulator primitive objects into meshes and eventually a GLTF format file.

This virtual world conversion idea came from the Sirikata project that was done at Stanford over a decade ago. Any continued work would build on the many papers available on things like automatic LOD generation of scenes. Things to do.

Then RaguOS adds to OpenSimulator a WebSocket interface talking the Basil protocol and making available all the assets created by Loden. From this, Basil can display the region contents. Extended RaguOS’s operation with some authorization stuff (OpenSimulator module OSAuth) but haven’t gotten all the way to adding the JWT tokens that will eventually underly all the authentication/authorization.

This led me to wanting to put up a sample region that people could try out. Deployment got me working on Docker and the creation of images for consistant operation of OpenSimulator. The output of this has been a repackaging of OpenSimulator into a set of scripts to build and deploy OpenSimulator as Docker containers.

With all that done, I’ve become overwhelmed with the amount of work that still needs to be done. Also, the on-line virtual world environment has changed. Augmented reality is the Cool Thing at the moment. The latest incarnation of virtual world companies have grown, stumbled, and needed to regroup. The Big Guys (Facebook, Amazon, Microsoft, …) have started to deploy their hardware and infrastructures. So, my little hobby project seems a dead end. Like it will never have all the capabilities that the virtual world community needs and my dreams of bigger impact seem out of reach.

This means I will probably moth-ball the whole Herbal3d project, bring documentation up to date, complete the architecture ideas, and then move on.

What will ‘moving on’ look like? I’m thinking robotics. I expect to get back to ROS2, Docker deployment of node logic, and playing with things using transformer neural networks to build model predictive controllers. Will be fun.