1: \begin{definition}[admissible weights]

2:   %We will see in Theorem \ref{OptimalWeightPreservesAccuracy}

3:   %that enforcing positivity of $\U$ (or any condition implied by positivity

4:   %of $\U$) by linear damping of the deviation

5:   %from the cell average preserves accuracy.

6:   If $\MA\ge\MAopt$, then we say that $\MA$ and $\WA$ are \defining{admissible}

7:   weights, because in this case positivity of $\U$ implies positivity of $\RAMA$,

8:   which means that enforcing positivity of $\RAMA$ respects accuracy.

9: \end{definition}

10: