That really depends on the methodology the software team employs. With Lean/Agile processes, quick, small, iterative releases are the defacto standard and recommended, especially when experimenting with new technology like torque sleep. The longer you take to put out a huge release, the chance for bugs/errors goes way up, plus you may spend a ton of time programming something that doesn't actually work or that the customer doesn't actually want, fail fast and iterate improvements to get more quickly to the optimal solution.
My guess is that they're testing two different solutions, one for .139, and one for .140, on P85D's to try to get real world data for a more optimal version coming soon. Ideally they'd want to look at how these versions perform in various places in the country based on temps, elevation, driver types, etc.