Alex Drummond

Ibex and the Ibex Farm

Ibex (formerly ‘webspr’) is a tool for running behavioral psycholinguistic experiments. It can do ordinary acceptability judgments, self-paced reading, and speeded acceptability. It's fairly easy to adapt for other things if you're comfortable with Javascript programming. Ibex itself is basically just a Python CGI script, and you can set it up on more or less any cheap hosting solution. Most of the time, it's easier to host your experiment on the Ibex Farm.

Numberize

Numberize is a Python script for automatically numbering examples and sections in ODT documents. I use it to make writing linguistics papers in google docs less painful. When you're done, you can download your document in ODT format and run the script to automatically number examples and headings. (The google docs export feature is a bit buggy, unfortunately, but the end result is usually pretty good for simple papers.)

Imask

Durham don't provide an email forwarding service or a POP server. This made it a bit difficult to get my dur.ac.uk email address forwarded to my gmail. Imask provides an ugly but workable solution to this problem. It repeatedly polls the Durham IMAP server, and then sets up a POP server for gmail to poll. (Gmail can be set to repeatedly fetch mail from a POP server, but not an IMAP server.) Imask is a node.js project available on github.

RJson

This was an experiment with the Scrap Your Boilerplate With Class paradigm in Haskell. The idea was to make serializing and parsing JSON date more-or-less automatic using data type reflection. In practice, the code ended up being extremely complex, and (in my opinion) a demonstration that SYBWC is not so great in practice as it is in theory. I think one or two people may have tried using it a couple of years ago, but it doesn't seem to have caught on. The module is still available on hackage.

Ant search

An attempt to replicate a simulation reported in this 1981 paper by Rüdiger Wehner and Mandyam Srinivasan. Desert ants perform a search when they fail to find their nests via their usual navigation methods. Wehner and Srinivasan argue that the ant's search pattern implements a near-optimal probabilistic search strategy. Source code is on github. You can also see the simulation here [non-IE browsers only].


a.d.drummond@gmail.com
Room 1.15a ArtsOne Building, 327 Mile End Road, London, E1 4NS