Yes Kai, the loops allow a factor to impact on itself, but it seemed difficult to understand for our partners. I was thinking a bit more about that this morning, and the way I ended up putting it was: if we change externally a factor x, it will impact the same factor x: for example if there is a spill (not in the model) in a lake, the "pollution" factor will increase, and this would impact "pollution"...
For the first part of my question, I used beforeValue() to implement a self-reinforcing loop (arrow: from is same as to). But now way to do this in qualitative mode (you could eventually modify the code so iModeler allows assigning an influence from a factor to itself, and add a curved arrow on the factor for display...
EDIT: ok I understand: I add a factor called "growth rate" which is say 0.1*population, which feeds back into population. Still the curved arrow would be nice