I don't know the sources... but (at my R1 university in math in the U.S.) many students from abroad do tend to use this. My comment on it is that it is all too easy to misread this. That is, it's not stable/robust from an information-theory point of view. And I do advocate writing a narrative (in natural language, e.g., English) rather than a parade of formal expressions without explanation other than formal justification. That is, I do not just want to see a computation, I want to first hear the plan of the computation/deduction, and comments along the way about how the specific low-level details fit into that plan... etc.
Sure, at an early stage, when people have just first gotten the idea of actual deduction and proof, being very atomic-detail-oriented is not a bad thing. But, as I also tell my students, especially grad students, at a certain point no one wants to see all the small steps, because they're willing to trust you and/or can fill them in by themselves. Instead, a more high-level/top-down narrative, that prescribes the atomic steps (modulo irrelevant details) replaces the line-by-line, eventually.