Kalusa was an ad-hoc collaborative constructed language, started by Gary Shannon and contributed to by maybe a dozen to two dozen people over the course of a few months in Summer 2006. The best description of the project I know if is by David J. Peterson, in giving the language his Smiley Award.
Here is the original source code for the Kalusa corpus management engine. It's a 40KB ZIP file containing 2629 lines of PHP and 780 lines of HTML and CSS. In email, Gary told me that he's released it into the public domain.
Below are data files saving sentences lost from the Kalusa corpus during an apparent automated-voting attack. Gary later modified the Kalusa software to disallow multiple votes on the same sentence from the same IP address.
The format of the files is tab-delimited, with four fields:
| Sentence number | Kalusa sentence | English gloss | CQ - correctness quotient, a weighted average of people's ranking of the sentence |
The original Kalusa software would, online, display the number of people who had voted on a sentence as well as its CQ, but this data was not included in the automatically generated corpora text files for download.
The Perl scripts I used for automating interaction with the Kalusa engine (e.g., downloading the corpus to a text file periodically, entering a batch of new sentences I'd composed offline, etc.)
Last updated September 2010
My conlang page