ABOUT
Easy-to-use open-source technologies that contribute towards the creation of documents in all the official South African languages.
WhAT IS AUTSHUMATO
Autshumato project is a first of its kind, open-source project that greatly benefits multilingualism in South Africa.
Initiated
Autshumato was initiated and funded in 2007 by the South African Department of Sports, Arts and Culture (DSAC).
We are currently busy with the 5th version of Autshumato and keep improving and adding to the languages available.
AIM OF PROJECT
To develop easy-to-use open-source technologies that contribute towards the creation of documents in all the official South African languages in the public sector.
Some of the outputs are software that simplifies the translation process, that promotes terminology standardisation and that shortens translation time.
DEVELOPED BY
These technologies are developed by the Centre for Text Technology (CTexT®) of the North-West University (NWU®) in South Africa.
CTexT is an innovative research and development centre at the NWU, Potchefstroom Campus. They run a variety of language technology projects, such as the development of free spelling checkers for ten South African languages.
Read more about them at their website
The latest version of the project is funded by South African Centre for Digital Language Resources (SADiLaR).
Interesting fact: Our project Autshumato’s name is derivative from the original South African translator, Autshumato. Autshumato was a chief of the Khoikhoi Gorinhaikonas who worked as an interpreter for the Europeans in present-day, Cape Town, South Africa prior and during the establishment of the Dutch settlement on the Cape of Good Hope in 1652. – Wikipedia
WHAT IS INCLUDED IN AUTSHUMATO
Machine translation happens when computerised systems attempt to automate the translation process from one human language into another. There is a difference between human-aided machine translation and machine-aided human translation:
- Machine-aided human translation happens when human translators use translation tools or software to help them with their translation work.
- Human-aided machine translation is when tools or software simulate the process of human translation with limited human intervention.
Translation memories are large databases of translated texts/units, which may include words, collocations, fixed expressions, and even short sentences. These units are activated (recalled or made available) automatically during the translation process. If relevant, the translator then merely must insert the unit into a text, in lieu of translating a whole piece of text from scratch. Translation memories are often used for the translation of technical documents, or documents with similarities to other documents. In the latter instances, translation memories help to avoid tedious repetition. Well-known systems working with translation memories include Trados®, DejaVu®, Wordfast® and OmegaT®.
Machine translation systems (a.k.a. ‘automatic translation software’) differ from translation memories in the sense that the translation process is fully automated, with no human intervention involved. An end-user usually pastes his/her text into the system, selects the language pairs, clicks on a button, and a full translation is provided forthwith. It is very important to note, however, that human intervention is required in text revision (‘post-editing’) of the machine translated results. Machine translation systems can be very inaccurate as context-sensitive translation through software is extremely difficult to accomplish. Technology can therefore never replace human translators, but rather contributes to an increase in the quantity and quality of translation work.
History & Future Projects
Our Autshumato Projects’ history and what we’re working towards.
The Autshumato Project was initiated by the Department of Sport, Arts and Culture
Added the English to Afrikaans MT system.
Added the English to Xitsonga MT system
First release of the Autshumato Translate App for Android
Added the English to Sesotho MT system
Fifth release of the ITE, all functionality is now implemented as plugins for OmegaT
First official release of the ITE
Added the English to isiZulu and English to Sepedi MT systems.
Launched the MT Web Services (MTWS) website
Fourth release of the ITE that includes the TMG and MTWS plugins for OmegaT
Added the English to Tshivenda and English to isiNdebele MT systems