TODO

Notes:
* Need to settle on how to do indexing/searching soon.
** Primary source of truth is HTML files on disk. Parsed format should be
   regenerated on the fly at server start time.
** Ideally don't want to reimplement indexing/searching.
** Data is highly relational. Relational DB might be best.

* Need to parse TceClass page contents into meaningful data, including:
** Normalization of different questions
** Normalization of different answer ranges, thanks to 2011-12

* Start bringing up a simple frontend. Thoughts:
** Try out Angular. Figure out what freemarker looks like for that.
** Bootstrap / etc.
** Add a PushStrategy to Jetty once we have a handle on needed resources

* Put together more request servlets:
** All of the courses in a department (with sorting qualifiers?)
** Details for a single course / class
** [Low-priority] add a LoggingHandlerWrapper for logging/tracing of requests. Make sure it
   does not block actual responses.

* TceClassModel:
* Add some simple calculation helpers for getting confidence interval and score

Bugs to resolve:
* Questions not parsed to enums
* Term strings not parsed to term codes (less important for now)
* Invalid pages (not logged in, load failures, etc.) shouldn't be saved/cached.