Similar numbers may measure different things
Two supplier tables can list 50 and 0.05 for what appears to be the same property, yet use different units. Another table may report a percentage while a neighboring document reports a fraction. Data cleaning research highlights representation inconsistencies across sources, and table interpretation research emphasizes semantic context. A comparison that ignores units can produce a false discrepancy or a false match even when every numeric value was extracted correctly.
Normalize with a visible rule
A proposed ABOLTUS comparison should preserve the original value and unit alongside any normalized representation. For example, a length expressed in millimeters can be compared with a meter value only after an explicit conversion. The interface should show the conversion rule and flag missing or ambiguous units for review. It should not infer a unit merely because nearby documents usually use one. The source's column header, table note or labeled field should provide the basis for the interpretation.
Test compatibility before arithmetic
An evaluation should include compatible units, incompatible dimensions and values whose units are absent. Reviewers should check whether the system refuses an invalid comparison instead of forcing every pair into one numeric scale. Different business meanings also matter: planned hours and billed hours share a unit but are not interchangeable. The intended ABOLTUS behavior is a comparison grounded in both measurement and meaning, with transparent transformations so a convenient normalized table does not hide assumptions that change the conclusion drawn from the original documents.