A place to publish research questions pursued with Scientific OS — not only the results, but the evidence behind them, how far the claims reach, and what remains unknown.
This is not a catalogue of finished work. Studies are at different stages, use different kinds of evidence, and remain open to revision through external reproduction and refutation.
Research in progress
GPU cluster scheduling
We study when running shorter jobs first helps, when it breaks, and which conditions decide the difference.
Queueing system identification
We test how much of a hidden queueing mechanism can be recovered from externally visible records alone.
Human movement model interfaces
We study how to stop coordinate mix-ups, target leakage, and false external validation before measured motion reaches a model.
Badminton biomechanics
We are designing an experiment that separates the effects of preparation time and centre of mass on the badminton smash.
The instrument in domains that require acting
What a seal-and-score apparatus can measure when the predictor is also the one who acts.
What a measurement design sees, and what it does not
How to doubt a measurement-design result that came out of a world you built yourself.
See current state by research area · Browse Research Notes by date