Begining with resources and their representations in or as files, on a variety of filesystems, on a variety of media, we want, in ascending order…

  1. A graph describing the files (e.g. path, location, dates, mime types) and resources (URIs) for which they provide a representation
  2. A static Apache HTTPD config generated from the graph to serve the files (or redirects) from resource URIs
  3. Some webserver implementation to generate a variety (file, styled, contexted, with navigation, etc.) of resource representations
  4. An evolving system (organic flexibility, feedback, community contributions, etc.)

Let's see, what to do for level one…

  1. Figure out the look of the graph model
  2. Find existing libraries to crawl an extract metadata from a filesystem
  3. First dump to a monolithic file, then think about a repository and automated consistancy (the graph is often a cache of the facts contains in the filesystem)
  4. Code