Core S² Software Solutions

Over the summer of 2011 I partnered with aXelerate Solutions, inc. to implement a simple proof of concept application for the iPad (iOS 4.3) to demonstrate it’s ability and usability for on-the-field applications. Specifically I wrote a POC for a local county’s department of property assessment; every county throughout our country has a group of assessors dedicated to keeping tax records and estimated property values over time. Though the current system is functional, it doesn’t allow data entry on the spot nor does it allow assessors to pull up important information (i.e. history of the property, terrain information, official zonings, etc..) at the location; most estimations are later done once the assessor is back in the office. In this video, I explain in more detail what this application does, what I have proven to be possible, as well as how the application deals with issues of connectivity (this county, like any other county, doesn’t have perfect cellular coverage).

I did forget a critical feature that wasn’t shown in the video: this application is able to load and render on-screen cartography layers. When assessing a property, as mentioned in the video, certain map data can be helpful. This application is able to render such details as well as allows users to combine several layers together.

I’ve been planning and writing code for a new game I’m temporarily titling “DwarfCraft”. Much like my other dozen projects, this one is mainly to teach myself about different technologies and algorithms. This time around, I went ahead and created a 19-page white paper on the game itself as a way to demonstrate my ability to design games. This paper is a far cry from anything that could actually be produced, but I think is a great starting point!

Along with this paper, I’ve been reviewing my old Cubism code that I wrote a few years ago and wanted to see if I could try again but solve the big hurdles I hit the first time through: memory usage and low frames per second. My current approach is to use an Octree data structure to represent my game geometry. This allows me to only allocate data that is actually needed for the game rather than the entire world. Data that represents empty space simply isn’t allocated, which dramatically reduces overall memory usage. Using XCode’s memory observer and leaks detection tools, I was able to prove that large worlds (i.e. 512 x 512 x 512 descrete cubes at one or two bytes per cube) barely takes more than 10 MB. Also, to help improve the FPS, I’ve started learning about VBO (Vertex Buffer Objects) instead of using the slow immediate-mode approach. I explain in much more detail in my video:

DwarfCraft White Paper

 If you’re interested in this project, make sure to contact me!

Two months ago I used to live off-campus near Penn State’s main University Park campus. My apartment was full of art students, with myself being the only non-art student. I decided to combine my knowledge with my roommate’s experience to create a real-time demo project using an Arduino and Gameduino. In essence my goal was to make something similar to a demo (read more via Wikipedia) fulfilling the requirements of: a unique graphics output, all computed at real-time, using only the Arduino processor and the Gameduino VGA output of 400×300 pixels. This project was quite a challenge as I was using an Arduino Uno which comes with one ATmega328: 16Mhz embedded processor with 2KB ram and 32KB programming memory!

Here is my result, recorded via camera (sorry, I had no in-line recording system for a VGA cable) and uploaded to YouTube. The demo lasts just short of two minutes. The final code compiles to about 27218 Bytes (~26KB) and includes a dozen effects: plasma, nyan cats, fun spirals, 3D cubes, pseudo cellular automaton, and a Portal 2 reference.

Just to make things clear: this entire video was generated by an Arduino and Gameduino connected via VGA to a standard TV screen; nothing was precomputed or loaded upon run-time, everything you see is done real time. 

Here is the Arduino code as a zip; make sure you are using at least an Arduino UNO as any other older device might not have enough ram to execute the code (and will have odd timing effects).