Apertium was developed as one of the open-source MT systems which originated within the project “Open-Source Machine Translation for the Languages of Spain” (“Traducción automática de código abierto para las lenguas del estado español”) (Forcada et al. 2010). The first paper that describes Apertium as a platform was published in 2006, although the origins of the system date back earlier than that (Ramírez-Sánchez, 2006). Apertium is an MT system designed originally for related languages, but later it added tools to facilitate unrelated language MT.
Even though the MT system described in this thesis only uses the shallow-transfer engine of Apertium, the platform is a full-fledged system for doing MT. It is at its core an analysis-transfer-synthesis system. The shallow-transfer engine is described in Section 5.9.2, but the other parts of the full system include finite-state transducers for lexical processing which are used for both analysis and synthesis, hidden Markov models for part-of-speech tagging and finite-state-based chunking for structural transfer. The system is fast, boasting a throughput of tens of thousands of words per second on an ordinary desktop computer (Forcada et al. 2010). Apertium also has a web-based front end (Apertium, 2015). From this website, you can translate texts or documents in 40 language pairs. Since Apertium is an open-source platform, new language pairs can be added by anyone at anytime. There are stages a language pair goes through until it is deemed mature enough to be called stable and therefore available on the website. Besides the 40 stable pairs, there are many pairs in various stages of development. There is quite a lot of documentation on the website and in Forcada et al. 2010, but I found the most useful help in the live IRC chat page which is manned by experienced users. It is here that practical problems can be solved.