jtolds.com - Projects http://www.jtolds.com/projects/rss.xml JT Olds' RSS Feed for Projects en-us Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License <![CDATA[Pants]]> JT Olds 1325650212 Wed, 4 Jan 2012 04:10:12 +0000 Over a year ago, for fun, I started creating a programming language. My favorite part of it, hands down, is the name.

Pants is a simple, stackless, dynamically typed, compiled, multiparadigm, statically but optionally hygienically dynamically scoped (holy adverbs), garbage-collected, continuation-based language. MLIA.

More information on the official Pants website.

Project links
]]> <![CDATA[Scheme2CPS2C4Py]]> JT Olds 1283447692 Thu, 2 Sep 2010 17:14:52 +0000 For CS6470 at cs.utah.edu, I wrote a simple Scheme-to-C compiler in Python, with Scheme in CPS as an intermediary.

I've wanted to put this up for a while, but I persuaded myself that at some point I would clean it up, write documentation, and add more features. However, neither is happened, so I'm cutting my losses and just uploading this project.

Project links
]]> <![CDATA[Dual N Back]]> JT Olds 1269847510 Mon, 29 Mar 2010 07:25:10 +0000 I've always had a bad short-term memory; it's atrocious. For most of my life, I've gotten by with rough heuristics and systems I've developed for maintaining some semblence of functionality. Recently, however, my inadequacy in recollection has begun to take more and more of a toll.

In 2008, some groundbreaking research was presented in the Proceedings of the National Acadamy of Sciences. This research suggested that not only was our understanding of fluid intelligence being somehow fixed at birth wrong, fluid intelligence and working memory share many of the same cognitive structures, and can be improved with simple memory games.

One such game, the so-called dual n-back continuous performance task, is simple, straightforward, and can be implemented for Android in a weekend, with time to spare for skiing. :)

This is my first Android application, and moreover, my first non-academic Java application, so it sucks hardcore, and it sucks more than just the standard amount of suck that Java brings into the equation. But it seems to work okay.

This is also my first GitHub submission, and it appears GitHub has decided to hate my mime-types, so you may have to right-click and select "save as" to download the APK or something screwy. Who knows.

This app requires Android 1.6 or newer due to the use of the built-in voice synthesis API for the auditory memory task.

Project links Dependencies
]]> <![CDATA[SHA1 Javascript Tool]]> JT Olds 1238267871 Sat, 28 Mar 2009 19:17:51 +0000 Occasionally, I need a SHA1 hash of something, and I'm not in a convenient place to get one. For instance, it is hard to open up a terminal to run sha1sum on public internet kiosks. It's also obnoxious to have to SSH anywhere from my phone to get access to a SHA1 algorithm.

So, I took Paul Johnston's Javascript implementation and made a simple webform. Trivial, I know. But I took out unnecessary HMAC calculation code, so it should be relatively easy to examine the source if you're worried about using something secure. The webdesign is, hah, optimized for mobile devices.

Project links
]]> <![CDATA[PyNarcissus]]> JT Olds 1235103926 Fri, 20 Feb 2009 04:25:26 +0000 So, I haven't actually just been watching TV recently, as my newsletter might currently suggest. I'm trying to be productive in a few ways, and one of them is the release of PyNarcissus here.

PyNarcissus is a direct port of the Narcissus Javascript engine by Brendan Eich. Narcissus is Javascript implemented in Javascript, and, while the evaluator does a few things metacircularly, the parser was pretty much a piece of cake to write in Python.

PyNarcissus does not introduce any new features whatsoever. Correctness using the original Narcissus as a baseline was an important priority, so even the pretty-printed string-converted output of the PyNarcissus parse tree is identical to the Javascript version. Using this feature, I simply ran both parsers on all of the Spidermonkey Javascript tests to make sure that the output was the same.

Because the port was so easy, I am considering going all the way and making a full on Javascript interpreter, but that wasn't my main goal for this port, so I'm not sure I'll get there soon, if at all.

Project links Dependencies
]]> <![CDATA[Viricide]]> JT Olds 1181979973 Sat, 16 Jun 2007 07:46:13 +0000 I'm currently working on a clone of a falling-pieces, Tetris-style game. My version is a networked implementation of a popular game that often features an Italian plumber masquerading as a physician.

I've been working on this project in various respects for over a year, as it was originally a way to teach myself Python. However, the project has since matured a bit and now features a client/server architecture, among other things.

I've open-sourced the project and put it up for development at Google Code. For more information, please see the official Viricide website.

Project links Dependencies
  • Pygame
    A Python library that eases the use of videogame creation.
  • Python
    The Python programming language is required to interpret Python code.
]]> <![CDATA[Horseshoe Implementation]]> JT Olds 1176592953 Sat, 14 Apr 2007 23:22:33 +0000 A while back I wrote about a game I was thinking about that I called 'horseshoe'. You can find more information on its entry on my newsletter. I'm considering researching various aspects of the game more, including strategies and such, and it occurred to me that having an environment in which to write artificial intelligences for the game would be incredibly useful. I suspect that it is a solvable game given a significant amount of computing power, but to show that I would need to actually solve it.
To facilitate this, I've written a very bare-bones Python game of Horseshoe. There is now an AI. It's pretty naive, but you can now play against your computer. In the immortal words of Donald Knuth, I have only proved it correct, not tried it. I think its win detection is right, which is the part that matters.
To run it, execute main.py.

Updated to include a reference to my poster I presented on Horseshoe.

Project links Dependencies
  • Python
    The Python programming language is required to interpret Python code.
  • Pygame
    A Python library that eases the use of videogame creation.
  • VPython
    A Python library that makes 3D programming really easy.
]]>