Richard Feynman explained that for some types of work, e.g. doing theoretical physics derivations, he needed large chunks of uninterrupted time. I personally think that any research work ultimately benefits from this kind of distraction-free workflow.
The reality for myself and, I venture a guess, for most of my colleagues (ok, for most people), is far from ideal, though. I do my academic writing in what Robert Boice calls brief, regular sessions (BRSs), and strive to keep them as regular as possible, without being concerned that they are more beef than I would like.
Likewise, I draw and paint on my iPad is sessions so short, that they cannot really be called sessions. Still, this ability to steal a minute or two here and there to do a sketch based on a photo that I took in Venice six years ago is precious to me. Sketching on a tablet is not perfect. It would have been nice to fire up Photoshop and draw with a Wacom graphics tablet, but the reality is that other obligations (many of them self-imposed or even imaginary, but this is beside the point) are so numerous, that long, concentrated painting sessions simply don’t happen, or at least don’t happen often enough
Working in microscopic, fragmented slices of time is a compromise, but it is better than not working at all. In fact, this fragmented workflow even has it’s unique advantages, but the main benefit is that it enables consistency, which is crucial for skill development, or improvement in general.