It's hard to keep track of which scripts are doing what in the sequence of transformations described in the previous post. Preparatory to implementing in XProc, I've coded up the pipeline in my own dataflow XML, which I transform to Graphviz dot and hence to an SVG image. This image is clickable so that the inputs, outputs and the intermediate scripts can be viewed.
Areas ranked by student density (cached)
I must switch to XProc notation before XML Prague or I'll have to hang my head in shame.