Some Things I Built...

Pirates Are BLANKing Awesome Link to heading

500 years Act 1 Link to heading

  • A janky game I made to play with open source; there will not be an act 2
  • Successfully marketed and was accepted in Steam Greenlight
    • ~35,000 units sold
      • Many more copies pirated :)
    • ~1,500 peak daily players
      • Potentially card farmers
  • Windows (Android & iOS variants were also released, but no longer available)
  • Built with MoaiSDK
    • Game in Lua, engine customizations in C++
      • Customizations include:
        • Including x-input
        • Glut modifications
        • Windows host modifications
        • Mobile host modifications (iOS & Android)
        • Steamworks integration
  • ~2 years to complete
  • Steam: https://store.steampowered.com/app/357200/500_Years_Act_1/
  • Game Source: https://github.com/jyakimischak/500years1/

Baby’s First Database (bb1db) Link to heading

BB1DB Design

  • A pure JavaScript database I made. Sometimes it’s good to build something from scratch to learn about it
  • Data is stored in Red Black trees, grammar and lexer are done using ANTLR
  • Table creation and inserts all work, got bored when I started on selects /shrug
  • Source: https://github.com/jyakimischak/bb1db

hl7x - Domain Specific Language For Hl7 Translations Link to heading

Hl7x Design

  • I got the idea in my head at some point that HL7 translations would benefit from a DSL
    • Specifically HL7 v2, “pipe and hat” formatted messages
  • This is a proof of concept for that HL7 translation DSL
    • Called Hl7x as “x” is often used as an abbreviation for translate
  • This:
    • Defines a grammar and lexicon for the DSL using ANTRL
    • Provides a multi-pass parser (written in Java) to read in scripts and run translations
    • Initializes itself with an HL7 xml dump of all 2.x versions
      • This file is provided by HL7
  • This is a library I wrote, mostly on my own time, but as it was shared with work I will not include the source code

Neural Net Link to heading

Neural Network Design

  • A toy neural network written in Go; building from scratch to learn
  • This implements a forward pass neural network
  • It’s got tests, etc, but I did not add in anything to train it, so all it can do is have a random “thought” right now :)
  • Source: https://github.com/jyakimischak/neuralnet

A Little PACS (Student Research Project From Long Ago) Link to heading

PACS

  • This was a pure Java Picture Archiving and Communication System (PACS) that I built for our research lab…going on memory here, I no longer have access to the source
  • It included:
    • Pure Java based on J2EE design principles
    • Custom DICOM network layer
      • Run as static code to be optimized for a single processor, we optimize things differently now :)
    • Custom DICOM header parser
    • DB2 storage with xml columns for meta-data
    • Used only Implicit VR Little Endian transfer syntax (good enough for a research tool)
    • Threaded GUI in Swing
  • Tested against the St. Boniface Hospital PACS and eFilm DICOM viewer successfully

PACS Performance Monitor (Student Research Project From Long Ago) Link to heading

PACS Perfmon

  • This is a performance monitor for PACS networks…going on memory here, I no longer have access to the source
  • This tool is much like other performance monitoring tools but uses the DICOM network layer of the PACS I created to send data
  • Studies of known size are sent from a network of test devices
    • These are generated using a pre-determined distribution (random intervals, set intervals, Poisson)
    • Studies are sent and data is recorded
    • The master node then compiles statistics and outputs a report