Setting standards for more effective courses information management
Generally speaking, an XCRI-CAP feed should enable an aggregator to collect all courses in a feed, or just updated data through small update feeds or filtering by date.
How often you publish your feed depends also on how often your courses change. There are different models of exposing your data. For instance, you may produce the XML from a data table which is updated at a set time by a batch process - and any systems can pull the up-to-date information from the re-created feed. There is a date/time stamp element within XCRI-CAP which indicates when the feed was generated.
Work is under way through JISC to determine how to implement 'delta harvesting' (just taking changed data rather than whole data sets).