FoundCAT - the new open source immersion resource

An Update on Developments

An Update on Developments

by Gary Evans -
Number of replies: 0

The FoundCAT Project

FoundCAT is an ongoing project at the FH Aachen university of Applied Sciences, which is based on the former Duolingo Immersion website for helping language learners improve their reading and writing skills though crowd-sourced translating.

The FoundCAT project at the FH Aachen aims to make use of computer-aided translation (CAT) tools to help students improve their writing while learning a foreign language. Gary Evans, FoundCAT project manager explains:

"We have built up a lot of experience and data including translation memories and termbases while translating using Memsource. Our aim now is to make use of these tools and the data to help students translate into foreign languages (L2), something which translators do not normally do as they traditionally translate into their native (L1) language.
We are also making use of machine translation to help users learn the structure of the language they are learning. After translating a segment, users can compare this to the machine translation and other users can make changes to it, as well as posting comments and links as feedback. 

Useful feedback is provided to the learner by highlighting errors and identifying the error type. This iterative process helps them learn to improve their writing skills as it not only corrects them and improves the segment, it also enables FoundCAT to provide the learner with links to lessons aimed at improving their skills.

There are many advantages with this project. Users can upload a text which is of interest to their studies and get it translated in the crowd. Here at the FH Aachen we are regularly supplied with press releases from the press office and the various faculties, which detail interesting projects and events. By translating these, we become better informed about current affairs and also we get to supply the press office with translations of the original text in return.

We're not looking for perfect and timely translations, we're interested in analysing the process and helping users learn from it. Our aim is to make FoundCAT available to everyone, so that users can learn from each other while translating creative commons content such as Wikipedia. 

Looking ahead, we hope to provide FoundCAT to refugees to help them learn German for example by translating from Arabic. This platform is about crowd-sourced learning and any language combination is possible."

FoundCAT is released under the GNU General Public License GPLv3. Please contact us at if you have any questions, or to receive access to the source code of our unreleased work in progress. FoundCAT is powered by Weblate and other free software. We are planning to add an LTI (Learning Tools Interoperability) interface soon to enable direct access to students of all universities (using any learning management system or LMS such as ILIAS or Moodle) without needing to register an account.

We hope to open up registration to provide access to FoundCAT soon. We'll keep you posted!