Parse the Matlab header files and build a tree to be able to perform
easy transformation. Having a tree should help at applying different type
of transformation (depending on the context) on the different section
(Example section, See Also section). But building a tree may require a
lot of memory.
Code example Section: We should be able to extract the example code
from the documentation, run separately those pieces of code and include
the textual output as well as the figures in the final documentation.
There should be a line-numbering system of the code and an automatic
labelling of the code and the figure.