How  does the whole thing work?

dataflow schema

Ok, now let's try to explain the workflow of the entire system. After a log in, a tutor can design his own exercise. Once he/she has finished, the exercise will be converted into an XML document and sent (inside an HTTP request) to the Zope server where ZoPark is running. The exercise can then be stored inside the Zope server itself, or inside an RDBMS, we simply don't care about this in this document.
When the exercise is used, an exercise framework is loaded and, when the users logs in, he/she is shown a selection of the available exercises. Once the exercise has been selected, its exercise definition is loaded from the server. Parsing this documents builds a list of the components needed by the exercise. These components are requested and loaded from the Zope server via the standard HTTP protocol. Dynamic loading of components is a feature of Flash MX. Once the exercise is over, another XML document containing results (answers, timings, etc.) is produced and sent to the server, to be stored for statistical analisys.