A researcher is never happy with a paper when it gets submitted.

Research never goes as fast as planned. Be honest about that, and people will understand.

When you have a thesis/project deadline, do not put other people’s work on your critical path.

Making new mistakes is progress.

Identifying what you or other people should have done or should do is progress.

Research often occurs in phases:

1. 1.

   have an idea

2. 2.

   work out the idea properly

Then it continues along one of the following two branches.

Theory:

1. 1.

   develop the theorems about your worked-out idea and prove them

2. 2.

   write a paper about your theorems

3. 3.

   use your theorems to solve problems

Practice:

1. 1.

   implement the worked-out idea

2. 2.

   make the implementation usable for others

3. 3.

   use the implementation to do a case study

Be aware of the following problem

• •

  each phase takes orders of magnitude more work than the previous one, therefore: moving to phase n+1 without fully understanding phase n is an expensive mistake,

• •

  each phase discloses problems that were invisible in the previous phase, therefore: you can’t understand phase n without moving to phase n+1,

• •

  good research constantly loops between the phases so that experiences from the later phases feed back into earlier phases.

In a PhD thesis, it’s quite normal that half your time is spent redoing earlier phases (including renewing your understanding of what the idea actually is).

Always do the simplest possible thing all the way through. Then try to do more general problems all the way through. If you try to do the first step for a general case, you will usually get stuck; even if you don’t get stuck, your solution for the first step will usually be wrong.

Two metaphors related to this lesson:

• •

  The tunnel metaphor: To build a tunnel through a mountain, first walk around the mountain (  manual input-output example) to see whether you want to be there at all; then build the simplest, smallest possible tunnel that you can crawl through (  prototype), then widen the tunnel (  development).

• •

  The pyramid metaphor: To construct a hiearchic building (  computer system), start with the smallest base area such that the tip of the pyramid over it reaches the intended height; then widen the base area turning the pyramid into a frustum with a widening plateau area (  the provided interface layer).

And an opposite advice: Sometimes foresight commands substantial changes right at the beginning. Metaphorically, we might foresee a much larger eventual pyramid frustum, too large to be supported by the chosen construction site; then we should start building elsewhere right away (or be aware that we first build a small practice pyramid before starting with the real one). Applying this foresight in the right situations requires a lot of intuition, expertise, and experience.