Potential models using XCRI-CAP - data collecting organisations
Data collecting organisations are usually interested in collecting either data on all courses or data on courses which have changed, or both. In a typical example there will be an annual data collection exercise to gather all the information for the new academic year, and the possibility of periodic or continuous updates during the year. The timing and frequency of data collection exercises is often dependent on the requirements of paper publishing, which tends to have more time sensitive deadlines than web publishing.
Some early work has been done on possible processes between learning providers and data collecting organisations, involving both ‘push’ and ‘pull’ models, but definitive examples using XCRI-CAP do not yet exist. In the ‘push’ model, the learning provider publishes the data and pro-actively sends it to the data collecting organisation using an automatic process whenever triggered by an update in the provider’s system. In the ‘pull’ model, the learning provider publishes the data on a passive service, and the data collecting organisation harvests it either on an ad hoc or routine basis. Both models can involve access control for security purposes and service level agreements.
The following diagrams show some examples of potential practice.
FIGURE 5: UML ACTIVITY DIAGRAM SHOWING A SIMPLE GENERIC OVERVIEW OF DATA HARVESTING USING XCRI-CAP
The collection of whole data sets is not problematic from the data collecting organisation’s perspective. Once the information is published by the learning provider, the data collector can scrap the old material and replace it with the new. This process is well suited to annual data collection.
The process for collecting updated information has additional constraints. Most data collecting organisations use quality standards for the data content, which may include both specific text formats for descriptive information (for example, ‘course titles will be in CAPS starting with the qualification awarded on completion’) and prescribed vocabularies for searchable fields (for example, the learndirect Classification System (LDCS) for subjects within the national Courses Directory. If the whole data set was to be used for updating purposes, then the organisation would either have to re-format and re-code all the data in accordance with its quality standards, a prohibitively expensive operation for intermediate updates, or carry out a complex comparison of old and new data, which might be technically difficult.
XCRI-CAP has specific features designed to aid the update process, for example the ‘recstatus’ attribute, which permit identification of new, updated and deleted elements. An update process using this feature might be as follows:
FIGURE 6: UML ACTIVITY DIAGRAMS, SHOWING UPDATING MODEL
A further model uses an open source data aggregating system called Course Exchange, which provides import, export, authoring and mapping functions, as well as enabling local data sources to be linked together. This system permits XCRI-CAP data to be imported and exported and may have the potential to enhance interoperability between disparate systems in different organisations.