In 1944, Soviet engineers receive three Boeing B-29 Superfortresses that emergency-landed on Soviet territory after missions over Japan. The USSR had no equivalent strategic bomber. Stalin orders an exact copy. Andrei Tupolev has two years.
The result is the Tupolev Tu-4. It enters service in 1947. Copy so faithful that the cockpit instruments are in American imperial units, that the fuel tanks have the exact capacity of the B-29 in gallons, that the bomb bay doors reproduce a slight asymmetry present in the original, probably an unintentional machining tolerance at Boeing. Tupolev did not copy the blueprints. He had none. He reconstructed the design constraints from the object.
Reverse engineering is not copying. It is archaeology.
Dismantling a technical object to understand how it was designed requires reconstituting the decisions that produced it: the materials available at the time, the accessible tools, the prevailing standards, the patents to circumvent, the accepted compromises. Each dimension, each tolerance, each material choice bears the trace of a constraint its designer encountered. The final object is the solution to a set of problems whose statement the reverse engineer must reconstitute.
What the reverse engineer reads is the space of impossibilities that their predecessor traversed. The absent forms of the object, the sections that would have been simpler to machine, the assemblies that would have reduced costs, the geometries that would have lightened the structure, are present in hollow. The object says what it is. It also says, for those who know how to read, everything it could not be.
Vincenti (1990) distinguishes in engineering knowledge the operational know-how, how to do, from the conceptual know-why, why thus and not otherwise. Reverse engineering reconstructs the know-why from the know-how crystallized in the object. It is the exact inversion of the design process: where the designer starts from an intention to arrive at an object, the reverse engineer starts from an object to arrive at an intention.
Doctrine
The technical object is the answer to questions it does not pose. To read an object is to reconstitute the constraints that made it necessary as it is rather than otherwise.
Vecteur ouvert
Tupolev was limited by the thermal resistance of available aluminum alloys, by the tolerances of his milling machines, by the absence of certain polymers that did not yet exist. These constraints are readable in the Tu-4 for those who know where to look. Our own technical objects bear the same imprints: the carbon cost that forbids certain geometries, the scarcity of lithium that constrains storage architectures, the simulation software that makes certain forms thinkable and others invisible because they do not fit within the mesh of the digital grid.
These constraints, we do not yet see them clearly because we are inside. The reverse engineer of 2150 will read them in our objects with the same evidence that we read Tupolev's limits in the Tu-4. The question is whether we can read our own blind spots before they become archives.
