The architecture

In addition to the various CaVa modules, there is a basilar component: the ontology, that serves to specify the knowledge domain that corresponds to the institutional information repositories. The actual input to CaVa (associated to module B) is the Learning Space Specification, written in CaVa^DSL, that determines which concepts should be exhibited and how they should be placed in the final virtual Learning Spaces by the Curator.

Module A, that is a kind of back office of this framework, is composed of the Database Repository and the Ingestion Function with the aid of a Document (Data) Ingestion System (DIS) to upload the institution documents information and data management. Module B is made up of four components that are CaVa^gen (a specification engine), CaVa^DSL Specification, CaVa^grammar and CaVa^grammar Processor, which processes the grammar and generates the specification engine. Module C is composed of the generated LS Scripts, the Web Browser, and the virtual Learning Space, which is the target or output of the CaVa System.

This architecture was not designed for a specific domain. Instead, it shall support any knowledge domain associated with museums, libraries, archival or schools.

Components of the architecture

Module A - CaVa^settler

The objective of CaVa^settler is to collect physical documents information from the institution Documents (the actual input of a DIS) and populate the digital database repository aiming at displaying the document's knowledge in virtual Learning Spaces. These are the components for this module:

• Ingestion Function (Document (Data) Ingestion System - DIS): it is the software (function) that performs the transport of the documents data to the database repository to be digitally stored. Basically, it extracts data from sources, storing the information into digital storage structures. Usually, this kind of system receives data through Optical Character Recognition (OCR) or forms that are manually filled;


• Database Repository: it is the digital storage structure that receives the data extracted (through a DIS) from the physical documents of organizations. These repositories can be classified as relational, object-oriented, triple store, etc.

Ontology

To generate the virtual Learning Spaces, which have the objective of imparting knowledge about a specific domain to the end-users, it is necessary to describe the knowledge implicit in the information contained in the sources (database repositories).

Depending on the Database Repository type, some extra tasks are needed (e.g. if the repository is a relational database, it is necessary to do the mapping between the ontology and the database, aiming at achieving the database instances).

Module B - CaVa^processor

It contains the main components necessary to achieve the objective of CaVa. CaVa^processor is the machinery that interprets the CaVa^DSL language specification (the CaVa input, manually written by the Curator) and uses CaVa^gen to produce as output, the virtual LS scripts necessary to render the desired final virtual Learning Space. Remembering that the input of this module is a specification written in the specific language CaVa^DSL. To be clear, these specifications are built manually by the Curator that should have knowledge about (1) the language rules defined by CaVa^grammar, and (2) the main ontology that describes the documents repository. The components of the CaVa^processor module are detailed below:

• CaVa^DSL: a Domain-Specific Language intended to aid Curators to create virtual Learning Spaces based on the museums assets and on the museological vocabulary;


• CaVa^grammar: a set of formal rules defining a formal language to be used to write sentences that specify a virtual exhibition. Phrases in CaVa^DSL are recognized by the CaVa^grammar Processor;


• CaVa^grammar Processor: a parser generator and a compiler constructor that recognizes the CaVa^grammar rules and generates CaVa^gen;


• CaVa^gen: a set of processors that performs two tasks before creating the LS Scripts: (1) recognize the CaVa^DSL input specification of the LS, written by the Curator; and (2) access the main ontology to search for the concepts referred in the specification in order to identify their instances.

Module C - CaVa^render

It is the module responsible for recognizing the LS Scripts generated by CaVa^gen and rendering, via a web browser, the virtual Learning Space described in a CaVa^DSL Specification. The components of the CaVa^render module are detailed below:

• LS Scripts: these generated scripts describe the intended virtual LS in a less abstract level, i.e. they are scripts that configure parts of the LS, like the static structure and content, dynamic content (queries), and the webpage presentation style and client-side files. The LS Scripts should perform two tasks: (1) query the database repository to fetch the real information concerning the ontological concept instances; and (2) use this information to create the final virtual Learning Spaces through a web-browser;


• Web Browser: an application for retrieving and presenting document's content (information) in the World Wide Web. In this work, the web browser has the duty of receiving the LS Scripts generated by CaVa^gen and rendering them, presenting the content of the exhibition rooms with the appropriate style, as described by the CaVa^DSL Specification.