Student projects 2018 edition


Here are a few hons/masters/PhD research project ideas. If you’re interested, then see the instructions on how to get in touch about doing research with me.

Seeing the bits: real-time visualisation & analysis of environmental data flows

Look around you—there’s digital data everywhere (this isn’t really a particularly insightful observation these days). There’s network data flowing over the wires and through the air, there’s data & computation embedded in many of the objects that you see, touch & ignore every day. And although these bits are largely invisible to you, they

Co-supervisor: Shane Magrath

Research questions

Required skills

Computing the assemblage

Co-supervisor: Tim Graham

Research questions

Required skills

In-browser livecoding with Extempore and wasm

Well, it looks like wasm is going to be a thing—a real thing. And since Extempore uses an LLVM backend for codegen it’s certainly possible

Research questions

Required skills

Live steering of scientific simulation codes

Co-supervisor: Henry Gardner

Research questions

Required skills

Learning to livecode

Co-supervisor: Henry Gardner

The jury’s still well and truly out on the best way to teach programming. However, livecoding—and especially musical livecoding—has a few things going for it as a vehicle for teaching

Research questions

Required skills

The Canvas: large-scale user-programmable data visualisation

Co-supervisor: Mitchell Whitelaw/Geoff Hinchcliffe

Code annotations for livecoding

Co-supervisor: Henry Gardner/Mitchell Whitelaw/Geoff Hinchcliffe

Teaching programming through musical (or at least artistic) coding

Why do we have to use java/python to teach programming? The reason I started doing this is because I wanted to make music, and the first programming that I did was musical livecoding in Scheme1.

The design & implementation of an interactive 3D graphics pipeline in Extempore

  • add animation to the graphics pipeline
  • canvas (hardware-accelerated charts/vis)
  • deferred rendering
  • general algorithmic music/DSP stuff
  • clean up the signal analysis stuff
  • ARM port
  • WASM port

  • doc update

Extempore on ARM: live programming the internet of things

Extempore2, an environment for 'live programming'3, has (experimental) support for ARM devices such as the Pandaboard, Hackberry and the ODROID-X2. These devices are capable of running mainstream Linux distros such as Ubuntu or Arch Linux, and support many different hardware sensors and actuators (e.g. LED lighting strips, servo motors for robots, etc.).

This project has three components:

  • building an ARM-based devices with on-board sensors & actuators
  • using Extempore, programming these ARM devices on-the-fly, including reading from sensors, processing data, and acting in the world.
  • contributing to the maturity of the Extempore compiler and runtime on ARM

Towards a toolkit for on-the-fly 'live visualisation' in scientific visualisation

Extempore4 supports the real-time programming5 of both 2D (with Cairo6/OpenVG7) and 3D graphics (with OpenGL8). The dynamic nature of extempore means that both the low-level (including vertex buffers and shaders) and high-level (including working with images, models, and lighting) components of the graphics stack can be constructed and manipulated by the programmer 'live'.

This opens up the possibility of 'live visualisation', where data (e.g. from a scientific computing application) is visualised in real-time in an exploratory fashion. The live programmer is free to modify the mapping from data to visual feedback on-the-fly to find the best visual representation for a given problem. This project involves investigating this type of live visualisation in Extempore, both from a computational and also a human factors perspective.

The goal of the project is to gain an understanding of what tools (both in Extempore and in a development environment) the live programmer needs to achieve this task, and the development of an Extempore library for live visualisation, and possibly a user study of this technique in a scientific visualisation context.

Coding for an audience: visual feedback and code markup for live coding performance

In 'live coding'9 a programmer/aritst writes a program to generate audiovisual material (often music) in real-time, in front of an audience. During a performance, the source code is projected onto a screen as it is edited for the audience's benefit.

This display of code for an audience is a key part of the practice of livecoding, but many audience members lack the required background knowledge (either musical, programming or both) to fully comprehend the relationship between the code on the screen and the musical material they are hearing. Some techniques have been proposed 10 for visual 'annotations' to the code which may assist the audience in their comprehension.

This project involves investigating various methods of presenting source code in the context of a live coding performance in Extempore11. This involves the design, implementation (Extempore supports both 2D graphics with Cairo12/OpenVG13 and 3D graphics with OpenGL14) and evaluating their effectiveness in live coding with a real audience.

Live data analysis/computational steering

Extempore15 is a programming environment for live programming. xtlang16, a LLVM-JIT-compiled programming language hosted by the Extempore compiler, supports toll-free linking and calling into shared C libraries (.so, .dylib or .dll depending on platform). This allows Extempore to use many open-source data analysis and numeric computation libraries, e.g. FANN for neural networks 17 or libsvm 18 for support vector machines.

However, Extempore's support for real-time JIT compilation (and re-compilation) and interaction by the programmer in a 'live programming' paradigm19 allows for the possibility of controlling these libraries at a high level, 'steering' them based on feedback on the computation in progress, or orchestrating multiple concurrent analyses of the same data, making decisions based on which ones were proving more fruitful.

This project would involve:

  • selecting a data analysis technique (or techniques) and a data set to analyse
  • implementing library/tooling support in Extempore to provide the programmer with appropriate feedback and control for the running analyses

Algorithmic music up and down the stack

Extempore20 is (among other things) a programming environment for algorithmic music composition21, and has built-in support a number of common tasks in this problem domain. These include several built in synthesisers and musical samplers (instruments) and audio/DSP effects (e.g. reverb, delay). Extempore, through it's strong temporal semantics22 also allows the scheduling of notes and other musical events in time, allowing for the possibility of generative music.

There are many different software environments for algorithmic musicmaking, designed for both offline and real-time use. What sets Extempore apart is the ability (through Extempore's support for both Scheme and xtlang23) to modify the generative processes at both an event (note) and at a signal (DSP) level, and all within the same programming environment. This project involves looking at algorithmic control and generation of musical material at both of these levels, including the possibilites of their interacting (e.g. note events manipulating the audio at a signal level, and vice versa). The outcomes of this project will be techniques for algorithmic composition which leverage this synergy, as well as a body of musical work which demonstrates these techniques in action.