# Calculation engin: consider “Acknowledge Circular Reference” in calculation engin.

Hi,

Basically, circular references are bad for the clarity of the model and bad for performence, and thus should be avoided where possible.

But there are cases, where the calculation is inherently circular, and can best be modeled using circular references. Of course, the circular formula should be “save”, e.g. having an exclusive calculation logic along the items of a category, so there is no circular calculation for any given cell. I call this a “save circular reference”. I know Quantrix can handle multiple iterations cell circular, but this is not what I’m talking about, as this is complicated use case, which I personally think should be avoided, but maybe there are justified cases even for this.

The topic here are the “save circular references”. There is a warning if you create one, and the possibility to acknowledge the warning. It seems that this ackowledgement does just suppress the error message. I think there should be a way to acknowledge and tell the calculation engin “this is a save recursion, I have made sure that never a cell depends on itself, you can optimise the calculation”.

In this way, the inherent performance problem of circular references could probably be solved.

To put it differently: I do not understand enough about the complicated calculation engine, but my feeling is that the two cases “I’m a dangerous circular reference, as a value depends on itself” and “I’m a save circular references, created by a gifted modeler which took care to make sure that nothing can go wrong” should be flagged differently and the engine should take advantage of it.

Thanks for considering.

Gilbert