The Traduco system
The Cnr-Istituto di Linguistica Computazionale “Antonio Zampolli” has developed Traduco, a state-of-the-art system for collaborative translation that facilitates, improves and speeds up the translation of complex works. In addition to the standard components typical of traditional CAT (Computer-Assisted Translation) tools, it includes features that simplify the translation process of texts with specific structural, stylistic, linguistic and hermeneutic problems.
The Traduco system, constantly updated, has been designed and implemented within the Babylonian Talmud Translation Project, which aims to create a collaborative Web system for the translation of the Babylonian Talmud into Italian. The definition of user requirements and specifications has been carried out in close collaboration with the project partners belonging to the Union of Italian Jewish Communities (UCEI) and the Italian Rabbinical College (CRI).
At the heart of the system is the translation memory, a specific component that uses a repertoire of all the sentences already translated and is able to automatically provide suggestions of possible translations, also on a semantic basis.
- Web-accessible: unlike desktop applications, it is used with a simple web browser and an Internet connection;
- collaborative: a translation team – consisting of translators, proofreaders and editors – can work on the same data simultaneously and interactively;
- intelligent: an assistance system automatically suggests the most appropriate translation for each sentence to optimise time and achieve more homogeneous translations;
- multimedia: in addition to textual resources, it is also possible to upload images, useful for advanced studies in the philological or palaeographic fields;
- adaptable to different languages: text analysis and processing components adapt to different languages due to their probabilistic nature;
- applicable to different domains: the tool easily adapts to domains of different natures;
- flexible and customisable: being based on a component architecture and open-source technologies, with minimal adaptations, developers can autonomously make extensions and customisations according to their specific needs (by language, target, type of media involved, etc.).
The system offers various functions; for example, it allows to:
- insert notes, comments, bibliographical references, annotations and citations;
- mark up the text on a semantic basis;
- automatically create thematic sections;
- perform complex searches within the original and translated text;
- read layouts created ad hoc and export translations for immediate production of printed editions;
- monitor in real time the progress of the work (percentage of completion of the text translation, work done by each individual user, etc.) via a supervision panel.