Harrison Nicholls / Software

Preface

This page indexes a number of computing projects I've developed by myself, mostly for fun. It is not an exhaustive list, and I've tried to keep each case quite brief. There's more information on each project on their dedicated pages, which are hyperlinked in each section below. I've also included a short list of software project ideas which I've yet to start on.

I've split up the projects into two categories: personal and research. The projects in each category are listed in chronological order, newest first. Projects marked with an asterisk* are considered complete. Projects marked with a dagger† are abandoned or are indefinitely on hold. Unmarked projects are under somewhat regular development. Projects are labelled with their year of inception.

Context

My first programming language was Java, which I learned at age 12; I used it to make some platformer video games with LWJGL (the same library that Minecraft uses). At around the same time I also learned some web design, which has not improved much since. At some point soon after I also picked up Python.

After dabbling with a variety of languages I eventually landed at C, which I fell in love with immediately. My favourite aspect of C is its low-level memory manipulation which provides a precision which few languages can come close to. It also has little computational overhead, and has a concise syntax (unlike Java). Structs are a super neat way to pass data around, and I've found that they provide a lightweight and flexible alternative to OOP. To date, C is my favourite language and I try to use it when I can. Sometimes I find that C++ is helpful as well.

Most astrophysical and atmospheric codes are written in Fortran 90. The performance of F90 comes close to that of C, although this also factors in the skill of the person doing the programming. There's a lot of legacy code written in Fortran hanging around, so I think it's a really helpful language to know. Physicsts and code can be a dangerous combination, but Fortran has made this precarious relationship reasonably manageable.

Python is incredible for data science. I use it for all my post-processing and plotting. There exists a vast array of libraries out there which can fill 95% of your needs. I would argue that Python is frequently applied in inappropriate contexts in the wild; I try to avoid adding to this problem by avoiding Python in many cases, although it can be helpful for prototyping. Dynamic typing is one of Python's most prominent features, but it's probably one that I dislike the most. Recently I've been moving towards Julia.

List of software

Wisteria

[2022] Scriptable parallelised standalone ray-tracing engine.

Wisteria renders scenes of objects to an image by projecting rays out of the camera and testing for their intersections with objects. At each intersection, the rays reflect, transmit, refract, scatter, and attenuate. Wisteria has the capacity to load objects from files on the disk. It's highly configurable and scriptable, although there's no formal documentation thus far. It supports distributed light sources and monte-carlo shadows & diffuse/specular reflections. It was developed for Linux, can also run on MacOS. There are example images and animations on the GitLab repository. It was originally prototyped in Python, but is now written in C++. Wisteria makes use of two optional libraries for file processing, but is otherwise standalone.

View Wisteria on GitLab

Website

[2023] Simple static website for indexing research, projects, etc.

I am not a web developer, which is likely evident from the site itself. I made the website for a number of reasons, the primary one being to index various projects and share them more easily. The secondary purpose of the website is to outline my research projects in a single location. The tertiary purpose of the website is to provide a profile of myself for others to read, if necessary. Written in HTML5 and CSS.

View Website source on GitHub

Planetary game

[2020]† Proof of concept game focused on the Martian landscape.

This was developed as a proof of concept. Inspired by Kim Stanley Robinson's "Mars" series, I wanted to develop some software which allowed you to interact with the Martian landscape across a range of spatial and temporal time-scales. It uses data from the MRO (NASA), splitting the suface into tiles such that the user can zoom in on regions of interest, while maintaining a high resolution. Ideally, the hydrology would respond to both rhelogical and atmospheric changes. This project has effectively been abandoned; some elements tie in closely with my research and are explored there much more scientifically. Written in C.

View Planetary on GitLab

Dripping tap

[2022]* Simple model of a dripping tap which can behave chaotically.

Toy model of a dripping tap which can exhibit strongly non-linear behaviour in certain flow regimes. This code is based on an existing model in the literature, although this implementation is entirely my own work. Like the logistic map, this model is a really beautiful example of a simple system exhibiting complex behaviour in certain regions of parameter space. As this code was developed for an undergraduate project, there's a short report in the GitLab repository which validates the model and analyses its behaviour. It'd be interesting to benchmark this against a hydrodynamical code, or even a real tap. Written in C, and depends heavily on the GNU Scientific Library.

View Chaotic Dripping on GitLab

Magic Printer

[2021]* GUI for producing proxy cards for Magic: The Gathering.

'Magic: The Gathering' is a table-top card game where people compete against each other using decks of various cards. The cards usually cost money, but by making fake versions ("proxies"), you can play the game without spending much money. This practice is legal as long as the proxies are not sold.

This Python code processes a file (.txt) containing all the cards in your deck, and produces images of the cards required to play with such a deck. The GUI allows the user to customise their deck through various options and sliders. The resultant images are intended to be used with 'FreePrints', an Android app which lets you print images cheaply. Written in Python, using Qt for the interface.

View Magic Printer on GitLab

N-body integrator

[2021]* Velocity verlet integrator of N-body gravitational systems.

This code evolves a system of massive bodies over time, accounting for their mutual gravitational interactions, assuming newtonian physics. The integrator uses the velocity-verlet method to accurately step the model forward in time. It supports dynamic time-stepping, which is helpful in cases where the bodies are rapidly accelerating. There's also a Python script for animating the simulation output. It's possible to set-up the model with complex systems: it has been used to model the solar system, including the Sun, planets, asteroids, and moons. Written in C.

View N-body Project on GitLab

Conway's Game of Life

[2019]* Interactive implementation of the titular game written in C.

Conway's Game of Life is an algorithm which is played on a 2D cartesian grid. Each cell of the grid is in a boolean state (alive/dead), and evolves according to a set of rules. The game evolves iteratively, with the rules determining, based on its neighbours, if each cell changes its state. This is meant to mimic population dynamics, although it's famous for producing a large array of patterns and 'machines'.

This project was originally started out of my own interest, however the timing was fortunate enough that it was assigned as a final project for an undergraduate module mid-way through its development. This software implements the titular game in C, rendering it live to a GUI using OpenGL. It allows the user to save and load game states from the disk, as well as interact via keyboard and mouse.

Read about the game on Wikipedia
View Conway's Game of Life on GitLab