# Definition

According to Halstead, N and $\hat{N}$ differ by 10% and the correlation coefficient is close to 1. The relation between N (observed length) and $\hat{N}$ (estimated length) seems valid. Then Halstead uses this relation to evaluate other metrics.

$\hat{N} = n_1 * log_2 (n_1) + n2 * log_2 (n_2)$

where

• n1 is the number of distinct operators used in a function.
• n2 is the number of distinct operands used in a function.