Abstract:

This paper describes the architecture and implementation of a mobile learning platform, which is primarily designed to conducting assessments of primary school students and for monitoring their performance. The platform is designed to aid teachers in developing countries by providing a detailed analysis of students’ performance and facilitating comparisons of learning outcomes using graphical charts and geotagged results. The system is based on the RESTful architecture with minimal coupling between the various components. The front-end for this system is developed using the cross-platform framework called Cordova which enables multi-platform deployment on Android and iPhone mobile devices. The middleware is implemented using NodeJS and the Express engine. The back-end uses a NoSQL, JSON-native database called CouchDB for its master-master replication and multi-version concurrent control (MVCC) capabilities. The master-master replication allows teachers to independently run their mobile-based assessment systems on a mobile platform in the absence of Internet connectivity. Assessments and performance data is automatically synchronized between teachers’ mobile devices and a central server when Internet becomes available. This allows educational administrators to pull results and to push assessments to their teachers’ devices in a dynamic manner. Assessments in the system are modelled using a JSONized version of the IMS QTI Lite assessment standard. Assessments items are either chosen from the database or uploaded by the teacher as compressed IMS QTI Lite XML files, which are later extracted and parsed into a JSON object for native database storage using CouchDB. A notable feature of the system is the use of JSON objects in every aspect of internal processing. The platform also supports a generic structure to support multiple curricula using a specially designed data model derived from analyzing the national/regional curricula from Singapore, Canada, Australia, U.S. (Common Core), U.K. and Pakistan. The performance of the system has been evaluated by collecting statistics on mobile power consumption and scalability of the server to support multiple teachers and administrators. A test-bed of over 1,000 grade 6 mathematics questions was used to evaluate the performance of the system.

Keywords:

ResfFul, CouchDB, IMS QTI, Mobile Learning, Assessment.