Terminologization in computing — the same word in natural language vs. source code -- Greg Bryant

Left (teal): GloVe 840B neighbors — trained on web text, books, Wikipedia. The word's natural-language semantic neighborhood.  Right (purple): code2vec / CodeSearchNet / GitHub corpus neighbors — the same word as used in source code. Green nodes appear in both neighborhoods.
Terminologization is the deliberate recruitment of an ordinary word into a technical domain, where its meaning is fixed with formal precision. Unlike semantic drift (a slow, organic, unconscious shift), terminologization is abrupt and intentional — a language designer chose these words because they gestured at the intended construct. The programmer who writes try {} and also says "I'll try" carries both representations simultaneously. Note that this is very far downstream from the cognition entailed. fMRI studies on actual recruitment of regions in both contexts has not yet been done. That would treat computer programming as a natural science, when it is, at the moment, an engineering discipline that borrows from the formal and natural sciences.