"Flow time. During the analysis and design stages, software development is an ephemeral, conceptual activity. Like any conceptual activity, the quality of the work is dependent on the worker's ability to sustain a "flow state" — a relaxed state of total immersion in a problem that facilitates understanding of it and the generation of solutions for it (DeMarco and Lister 1987). Converting brain waves to computer software is a delicate process, and developers work best during the hours they spend in this state of effortless concentration. Developers require 15 minutes or more to enter a state of flow, which can then last many hours, until fatigue or interruption terminates it. If developers are interrupted every 11 minutes, they will likely never enter a flow state and will therefore be unlikely to ever reach their highest levels of productivity."
As I recall, in one of McConnell's books was a chart showing relative performance effectiveness of closed offices vs cubicles vs open floor plans, which got progressively worse. I can't now find that chart though.
The identification of flow state in programming was known well before Graham. Steve McConnell's Rapid Development (1996), https://archive.org/details/rapiddevelopment00mcco/page/506/... , has
"Flow time. During the analysis and design stages, software development is an ephemeral, conceptual activity. Like any conceptual activity, the quality of the work is dependent on the worker's ability to sustain a "flow state" — a relaxed state of total immersion in a problem that facilitates understanding of it and the generation of solutions for it (DeMarco and Lister 1987). Converting brain waves to computer software is a delicate process, and developers work best during the hours they spend in this state of effortless concentration. Developers require 15 minutes or more to enter a state of flow, which can then last many hours, until fatigue or interruption terminates it. If developers are interrupted every 11 minutes, they will likely never enter a flow state and will therefore be unlikely to ever reach their highest levels of productivity."
DeMarco and Lister 1987 is "Peopleware: Productive projects and teams". I can't find a first edition but you can read the relevant part of the second edition at https://archive.org/details/peoplewareproduc00dema_0/page/62... .
As I recall, in one of McConnell's books was a chart showing relative performance effectiveness of closed offices vs cubicles vs open floor plans, which got progressively worse. I can't now find that chart though.