Cloud Computing for Linguists

The system presented is a web application designed to aid linguistic research with data collection and online publishing. It is a service mainly for linguists and language experts working with language description of less-documented and less-resourced languages. When the central concern is in-depth linguistic analysis, maintaining and administering software can be a burden. Cloud computing offers an alternative. At present mainly used for archiving, we extend linguistic web applications to allow creation, search and storage of interlinear annotated texts. By combining a conceptually appealing online glosser with an SQL database and a wiki, we make the online publication of linguistic data an easy task also for noncomputationally oriented researchers. 1 General description of TypeCraft TypeCraft (or TC in short) is a multilingual online database of linguistically-annotated natural language texts, embedded in a collaboration and information tool. It is an online service which allows users (projects as well as individuals) to create, store and retrieve structured data of the kind mainly used in natural language research. In a system featuring graded access the user may create his own domain, invite others, as well as share his data with the public. The kernel of TypeCraft is morphological word level annotation in a relational database setting, wrapped into a wiki which is used as a communication and information gathering and sharing tool. TypeCraft allows the import of raw text for storage and annotation and export of annotated data to MS Word, OpenOffice.org, LATEX and XML. The online system is complemented by an offline client which is a Java application offering the same functionality as the online version. This allows a seamless exchange of data between the server and the user’s own computer. 2 Online system internals The online system is supported by a central server running the following modules: TypeCraft server proper, an SQL database, Apache, MediaWiki. The client side consists of the TypeCraft editor interface and a wiki environment (content produced by MediaWiki on the server). Users perceive the wiki and the editor interface as a single TypeCraft web application. The TypeCraft server proper is a Java application running inside a Java application server. TypeCraft uses a PostgreSQL database for data storage. The data mapping between Java objects and database tables is managed by Hibernate, so the system is not bound to any specific SQL database. TypeCraft data can be divided into two distinct groups: common data, shared between all annotated tokens and users, such as the word and sentence level tag sets and an ISO 639-3 specification, and individual data, by which we mean specific texts, phrases, words and morphemes. Individual data references common data types. This for example means that all users of the system making use of the part of speech tag N share the reference to a single common tag N. 3 Digital linguistic data It is well known that generation of linguistic annotation of any kind is a time consuming enterprise quite independent of the form the primary data has and the tools chosen for processing this data. Equally well known are problems connected to the generation and storage of linguistic data.