TaLTaC 3.0. a multi-level web platform for textual big data in the social sciences

Abstract

The TaLTaC software package as a tool of lexical and textual analysis, versions 1.0 e 2.0, lived over the last decades (1999–2015). It appears now to have met its technological limits. The TaLTaC version 3.0 (from now on T3) has been redesigned to overcome those limits. The process included: (i) recoding of all inner software components with modern web-related languages and standards; (ii) adoption of a new kind of database (NoSQL) capable to handle corpora in the order of magnitude of gigabytes; (iii) new criteria for data storage and data processing. The software architecture is modular and allows to decouple user interaction from actual data computing. The two main components are: the GUI (graphical user interface), based on HTML5/CSS/Js and the back-end processing CORE. The new design also made it possible to run T3 among the mainstream operating systems: Os X, Windows, and Linux. From a single parsing operation, T3 produces many vocabularies for multi-level lexical analysis. This allows one to disambiguate, in a semiautomatic fashion, between the different text graphical forms on the basis of concordance. I also allows for a virtual transformation of simple forms into multi-words.