We do a poor job at motivating the total error metric. Many reviewers respond viscerally to our long list of definitions.
This metric is essential for showing that CheckCell's error-reduction/effort is better than Normal-type analyses. We should defer the detailed explanation for after we've explained the metric's importance. Further, each sub-definition should be explained in the larger context: e.g., "this metric exists so that we can treat the results of string and numeric hypothesis tests homogeneously."
We do a poor job at motivating the total error metric. Many reviewers respond viscerally to our long list of definitions.
This metric is essential for showing that CheckCell's error-reduction/effort is better than Normal-type analyses. We should defer the detailed explanation for after we've explained the metric's importance. Further, each sub-definition should be explained in the larger context: e.g., "this metric exists so that we can treat the results of string and numeric hypothesis tests homogeneously."