The input data was files with one record per line, in a file tree. Files and directories had unpalatable names, so a special file 'NAMES' was created in each directory to translate to more meaningful names.
The scripts were written in the Python language.
One script translates the data file tree to a corresponding tree of HTML files. At any non-leaf node of the tree, a search (using glimpse) can be made, limited to the chosen subtree. The search result is then translated to nicer HTML, with pointers back to the relevant parts of the corresponding HTML tree.
Gary Capell (gary@cs.su.oz.au)