the self-similar growth regime (SSGR) and non equilibrium growth Questions tackled : \begin{itemize} \item Can we apply the tools of non-equilibrium statistical mechanics (including renormalisation group) to the SSGR ? \item Do all foams reach the, or a, SSGR ? \end{itemize} We should distinguish 2D foams from 3D foams, as well as dry from wet foams. However, two other cases where the coarsening, and the SSGR, should be modified are the flowing foam and the foam under drainage, where the gas doesn't pass from a bubble to its direct neighbor. As an example, what is known about the SSGR in the 3D dry foams case ? We expect the Faces distribution and normalized bubble size distribution to be stable with time, along with any other quantity once it's normalized by the average size $< R>(t)$. This seems to work for various initial bubble size distributions as Pott's model simulations have shown. What about the stability of the SSGR when it is perturbed? This last question leads to several remarks : whether the stable "variables" in the SSGR are enough to describe the perturbed SSGR and the transient regime towards the SSGR. What are the tools we can use ? What would be needed is a rate equation written in the space of normalized variables in order to describe the SSGR. This wouldn't imply that the right formulation be a mean-field theory and one could imagine using the Aboav-Weaire law for example. It's difficult to build a PDE theory based on discrete events (T1s). How many T1s, and are they correlated, in order to ensure the stability of the face distribution ? THink of an infinite foam? Regarding correlated T1s, T1 avalanches are observed in sheared foams - one could imagine the same kind of phenomenon taking place in foams in the SSGR. The frequency of T1s depends on the width of the size distribution of the bubbles. A wide distribution could enable the avalanches. Hence it might be interesting to simulate foams with different size distribution width in order to see if there is a critical width above which T1 avalanches are triggered. The question of the uniqueness of the SSGR state is then asked in the case of 3D dry foams. Results of Potts simulations seem to indicate that different initial distributions lead to the same SSGR state. The results of X-ray tomography experiments show the same feature although one could not say that the initial size distributions were as broadly distributed as in the Potts simulations. A regular feature of the transient regime towards SSGR : it seems that the average radius of the bubbles (for an unknown distribution) is multiplied by $\approx 3$ before the SSGR appears. This leads to an important loss of bubbles ($\approx 96\%$). the SSGR regime is reached in Potts simulations (dry and wet foams) when there are around 5000 bubbles left. What should be the competences needed to apply Renormalisation Group methods to the coarsening? it's a condensed matter theory problem and that there are two added difficulties to the usual RG problem : the disorder in foams, whereas RG methods are easier to apply in ordered meshes; and the combination of continuous and discrete variables. Difficulty to combine renormalisation with the topological disorder in foams. About the extent of mean-field theory attempts, the IPP mean-field theory precisely ignores the effect of the environment of a bubble, the growth of which is fixed by its own characteristics whereas the RG methods are usually introduced to take the environment into account. Could Self-organized criticality models be of any use? Identification - using simulations - of useful parameters to build a theory, e.g. constraint relaxation and redistribution. What about maximum entropy methods in order to characterize the SSGR distributions? this should enable to characterize an equilibrium state, not a growth regime. However, there are several refined maximum entropy methods, some of them based on the introduction of a free energy. None of those theories is devised to describe the coarsening and underlines the need for a normalized rate equation in order for the scaling regime to be described a an equilibrium state in the normalized phase space. What are the stationary quantities that can be measured. $n_{eff}$ in wet foam simulations; Minkowsky functionals might be good candidates. Surface Evolver simulations could evaluate the different functionals and their importance in the coarsening mechanism for foams experiencing a modified kind of coarsening.