code rot review

[slivingston]

Some parts of TuLiP are rarely or never used. In general having good test coverage can prevent rarely used code from unexpectedly breaking. Nonetheless, maintaining rarely used code as TuLiP evolves is not well motivated. However, because usage of TuLiP in the wild is not well known, it is difficult to decide which parts should receive priority of attention.

This issue is intended to track reviewing of the sourcetree as it is now, and

• to decide which parts should be deprecated, if any, and
• to discover if any parts are currently broken, despite tests passing.

For now and the future, the following protocol is proposed:

1. add deprecation notices to parts that we, as the developers, do not think are widely used.
2. create issue to indicate decision, to collect discussion, possibly change original plan of deprecation.
3. and then allow 1 year (or 6 months?) to pass before removal.