Points to talk about:
  * Presentation convergence profiles real engineering problems.
  * To choose: which direction ?
     - Present the three methods and say that on these types of problems, EB is the most efficient (but not the more robust).
     - Keep the accent on the progressive barrier.
  * Add or not the survey in the phd ?
  * Start the third project ?
  * Possibility to discuss to get the lexicographic objective application from Hydroquebec.
  * Fellowship : bonus.
  * Not really urgent : for the juliaopt optimization conference, will you like to take a look at the slides (it is free ad for your software) ?