Fable
The blacksmith merged three back-and-forth steps into one hammer strike
At the foot of the mountain, the blacksmith shop forged axles. The old process was very orderly: first heat, then carry from the forge to the anvil to hammer, then carry to the water trough to cool, then carry back to the anvil to finish. Each step was handled by a different apprentice, and on paper it was clearly divided.
When there were few axles, this was no problem. Later, the military camp needed three hundred at once, and apprentices spent all day carrying semi-finished products back and forth between the furnace, anvil, and water trough. The iron cooled before it was hammered, and the line jammed at the door.
The shopkeeper first hired more porters. With more people, the yard became more crowded, and the time hot iron wasted on the road did not decrease. The second patch was to make each process faster; the masters hammered harder, but it still could not overcome repeated moving.
The old blacksmith watched for a day and said the real waste was in the handoffs. He merged the consecutive small motions onto the same long anvil edge: the tapping point was right beside the furnace mouth, and after tapping he quenched the piece in water, then trimmed the edges in the residual heat. Whatever could be done in one pass was no longer split into three trips.
The new layout required retraining, and not all processes could be forcibly merged. Some large pieces still had to cool separately, and some fine finishing had to be inspected independently. But for axles suited to merging, speed rose immediately, and heat was retained better.
The blacksmith shop later understood that optimization is not just about making each step faster. If many small steps always occur consecutively on the same material, merging them into one flow can save the cost of back-and-forth transport, waiting and repeated reads and writes.
The old blacksmith also warned that merging has boundaries. If a crack needs to be checked between two actions, forcing them together will carry the defect to the next step; if materials have different temperature requirements, you cannot rush either. Merging must respect process dependencies.
After the army order ended, the blacksmith shop listed common axles, hinges and short nails as mergeable process steps. Each time they moved one less step, the yard got clogged one less time. The real gain came from reducing back-and-forth in the middle, not from simply writing all the process names together. Later, when the apprentices reviewed the work, the old blacksmith had them count not how many hammer blows were struck, but how many trips were saved and how many times the metal went cold. The gain from merging is hidden in the intermediate losses that are eliminated. The clearer the places suitable for merging, the faster the workshop can go without sacrificing quality. The apprentices also learned to look first for places with continuous, repeated, heavy handling, rather than blindly merging every step.