Datasource Module

The datasource module in our test is basically a wrapper around Microsoft's XML Parser(14). The main disadvantage of implementing a possible large datastructure in such a way lies in the poor processing speed. Using a closed source API such as the XML Parser adds the extra problem that it is uncertain where speedups can be implemented. During development different versions of the parser have been released. These all had several speedups over previous releases but just as many slowdowns due to increased standard compliance. More practical issues occurred when doing large scale stresstests such as excessive memory requirements. Looking back at the project I have to say that using a simple document as datastorage is far from efficient. An alternative that is probably the best choice for future use is an XML database. XML databases can probably overcome the practical issues I've encountered while allowing us to keep using XML as our primary communication method.

The format used to write file and plug-in data is modeled to the RDF. An RDF dictionary for standard file data called the Dublin Core is used to give meaning to items. Too few dictionaries currently exist to fully exploit the use of RDF at this moment. Therefore I haven't tried to fully comply to the RDF standard. I believe that developers should try to comply to the standard as soon as specialized development tools and generally accepted dictionaries emerge.