We’re proud of our fans.

There are two things that keep us working every day and night. On the one hand it’s the endless fun and delight of doing uber-awesome rocket science and on the other hand it’s YOU, our fans, who support and encourage us. We don’t want to miss any of the two side, neither the rocket science nor the fans and Part-Time Scientists wouldn’t exist any longer if you take away one.

And that’s the reason we want to say “Thank You!” Thank you, for being such great fans. Thank you for cheering us up. Thank you for supporting us every day since we entered into this race to the moon. And we’re confident we’ll win – with YOUR help!

YOU have given us so much, we want to give you back a bit. We came up with the idea of starting a small contest to celebrate all the 50.000 of you. We will start on facebook at the beginning of the week, more details will follow on Monday.

By the way: GREAT prizes from NVidia, 3Dconnexion and Part-Time Scientists are waiting for YOU!

“What engine/tool are you using [for 3D graphics]?”

And we promised to answer that question properly, so here’s what Wes, our Head of Software guy, says:

“What engine/tool are you using [for 3D graphics]?”

We use several technologies for our graphics and are currently consolidating our tool selections. To really make the choices, we have to understand why we’re building our applications.

A rover on every desktop

A critically important application is our simulator. When you’re testing AI logic to avoid a crater, code breakpoints and cout statements just don’t work as well as watching Asimov heroically dodge a crater. A 3D simulation engages our fellow team members and, frankly, engages us developers as the fun loving humans that we are. Our simulator is shaping up as C++-based OpenGL using Visualization Library, Qt, and Bullet Physics.

We’ve tried handwritten OpenGL in C++ and 3D in the Java-based Processing environment. Both were Ok but not quite what we need. Custom crafting OpenGL code is time consuming, so we selected a library to make terrain and 3D model visualization quicker to write without sacrificing the cross platform nature of OpenGL. 3D in Processing is easy to do but ultimately risky for us – it’s a common complaint when developers move from Java back to C++ that they forget to call delete and programs spring memory leaks.

Visualizations

Speaking of AI, we also see a need to peek into the inner workings of complex mission critical algorithms. Again, breakpoints and logging statements can only go so far. We can write complex behavior monitors to look for unusual behavior, but nothing, and I mean nothing, so far can beat the insight of a fully engaged human. We’ve used the Graphviz tool to look at genetic program output and are considering the Ubigraph software for real-time animation of complex data structures.

The future

Ultimately, our simulations and visualizations will find their way into our ground control software. It’ll be tied to GPU-based (CUDA and OpenCL) risk assessments and to Qt-based remote control software. Personally, I’d love to see the whole control room projected onto walls with everyone using augmented reality tablet computers. Framework suggestions and volunteers welcome for that one!

-Wes

• the slides and the actual code
• blogpost from Wes about his talk

You may also be interested in the list of all our talks at the 28C3 this year:

• Automatic Algorithm Invention with a GPU- by Wes
  • Recording of the talk
  • Further resources (slides and code) and you can rate the talk here and give feedback.
• Not your Grandfathers moon landing- by Robert and Karsten
  • Recording of the talk
  • No slides but you can still rate the talk here and give feedback.
• Evolving custom communication protocols- by Wes
  • Recording of the talk
  • Further resources (slides and code) and you can rate the talk here and give feedback.

• the slides and the actual code
• blogpost from Wes about his talk