Some Principles

• The repository can accept any type of content (might be restricted to known file types)
• XProc is used to process XML files: all parsing and transformations must go through XProc
• Different XProc pipelines can be activated depending on the action and the file type considered

Storage Specifications

• Dependencies (of XML files: XML, XSL, SVG, etc.) are parsed at upload time and stored (maybe along with the file) so that information can be easily reused for later purposes (display of a dependencies tree, prepare compilation, etc.)

Transformation specifications

• When a publication is requested, a compilation environment is built that contains all recursive dependencies of the XProc pipeline (that will notably include master XML and the XSL).