Downloaded from https://www.everyspec.com APPENDIX C MIL-STD-31000A model is to overlook the negligible variation inherent in the process which created the derivative and to clearly identify any significant variation which is unacceptable at that stage in the model’s lifecycle. The best practice is to match entities of the same type using key characteristics which generally have minimal variation introduced in the derivation process. For example, a graphic translation of an annotation should preserve the location, orientation, attachment point and total length of its presentation curves.
But a semantic translation, or manual remastering, will often only preserve its orientation and associated geometry – introducing significant variation in its location, attachment point and presentation curve length. Entity matching must account for common structural changes such as filtering, splitting and merging. For example, if a derivative model generation process always ignores hidden entities then these must be filtered by the validation process. If it is common for a cylindrical face which defines a hole to be split into two hemispherical faces, this must be accounted for during validation.
Entities in either model which have no acceptable match in the other model should be identified as unmatched and should not be included in deviation or difference calculations. Entities which match one-to-one, or in a split or merged relationship, should then be validated with the appropriate deviation and difference criteria. 67