This project has a long history, almost ten years now. I have been interested in communication technology, including translation, for many years. I have an interesting background and have worked on a number of projects directly and indirectly related to the worldwide lexicon.
1998-2001 : The first translation project I worked on, Picto, was a system that created something similar to the Domain Name System for words or concepts. In this system, each unique meaning for a word or phrase was given a unique numeric address, similar to a computer’s IP address, and was also associated with a pictograph or icon. The goal of the project was to create a standard way to embed these codes within documents, so that automatic translation systems deliver higher quality translations. In a web page, natural text would be annotated with hidden comment tags that referenced a word or phrase’s Picto code. While this system never took off, it has led to other inventions, such as the Disambiguation Markup system proposed on this website.
2001-present : SETI research : I have published one book and several peer reviewed articles about SETI and inter-species communication, and as part of my SETI work, have developed a general theory about how to communicate using mathematical languages and symbol sets. This work was heavily influenced by my work on Picto. I also gave a talk at the NASA Bioastronomy conference in Australia in 2002. I am currently working on another paper that describes how to compose self-descriptive messages.
2002-2003 : GNUtrans : This was our first attempt at building a hybrid automated/volunteer translation service for websites. Our goal was to build a system similar to SETI@Home, except that instead of tapping idle CPUs to do calculations, we would use instant messaging to ask volunteers to translate texts to and from various languages. We developed most of this system, but did not launch it commercially due to technical issues, mostly due to instant messaging interoperability. This work was quite valuable, and led directly to the simplified translation service we are launching this year.
2004-2005 : MIMS (Multilingual Instant Messaging System) : in this project, we prototyped an instant messaging system that would match bilingual IM users with people who wanted to communicate, and would bridge the translator into a conversation (similar to a three-way telephone call). We developed a lightweight Java IM client. We are planning to relaunch this service sometime in the next year, after we have studied the results from our new website translation project.
Present : we are currently preparing to relaunch a simplified version of our website and RSS translation service. This project is an experiment in “distributed human computing” where a large number of volunteers will contribute to translations for websites participating in our network. This project, dubbed TRIKI (translation wiki), is in many respects a dumbed down version of GNUtrans, and also embodiesmany lessons we learned from prior incarnations of the project.