I have so far developed the process for producing the XCRI-CAP feed. The application runs overnight to populate a separate set of tables for the XCRI-CAP datasource and then creates the main XCRI-CAP file. This has been validated using the JISC validator tool (http://validator.xcri.co.uk/ ) and passes with some warnings (largely on missing abstracts and descriptions). The application also produces a human readable version of the feed as an HTML page and an Excel file.
The main outstanding issues relate to some data discrepancies, some courses missing descriptions (which is allowed but with a warning) and some not having subjects. Those courses that do not have subjects must be resolved as they prevent a valid XML document from being produced. Currently these courses are excluded from the feed.
The application has been deployed to a development server. I am still waiting on the test and live environments to be made available as well as allowing the feed to be generated over SSL.