On a Good Thesis

• At the time of defense, a PhD should be the world's expert in something (cf. title of thesis).
  
• You can do good theory or good practice or both, but never use one as an excuse for a shoddy job on the other.
• People expect a PhD to be able to think long and hard about a subject---and say something precise and profound.
• Good ideas are easy, turning them into something substantial is hard (cf. previous point).
  Corollary: A good idea is not a thesis.
  Note: A good idea may be the seed for a thesis (cf. proposal).
• Where possible, a good thesis builds on the good work of others (cf. proposal, references).
• Creation starts with a shapeless void; it is an act of successive division (void from non-void, etc.) [cf. Genesis]
  Corollary (D. Knuth): Define and conquer.
• Everything is a tradeoff. PhD's have to make reasonable ones to make progress.
• A good thesis is a good argument: objective, reasoned, convincing.
• A good thesis tells a story.
• A good thesis becomes a reference book. Include a notation page, an index, appendices on background material, etc.
• People generally read in this order: title, abstract, introduction, conclusions, all the rest.
  Corollary: a good thesis has a good title, a good abstract, a good introduction, ...
• Portions of your work should be publishable in journal form.
• ``Future work'' is not an excuse for doing less than a thorough job in the first place.
• You should know more than what's on your slides.