Difference between reconstruction and discrepancy analysis

RECONSTRUCTION

DISCREPANCY ANALYSIS

Principal purpose

To understand the composition of a single calculation.

To identify differences between two calculations that are supposed to give the same result.

Input

A single cell, or a number of consecutive cells in a row that all contain the same formula

Two cells which contain results from the calculations to be compared

Output

Reconstruction reports at one or more levels of precedent, each on a separate worksheet

A fan-out covering (usually) several levels of precedent, laid out on the single worksheet.

Non-obvious use

Although OAK's principle weapon against inconsistent results from similar calculations is the discrepancy analyzer, sometimes useful insight is gained by reconstructing a cell containing the unwanted difference between the two calculations

OAK's reconstruction capability is valuable as a starting point for preparing a parallel reconstruction of a calculation.

But sometimes a discrepancy analysis is useful for the same purpose; since it shows on a single sheet contributors to the calculation that are at different levels of in the precedent hierarchy.  These can then be cut and pasted into the developing parallel reconstruction.