Figure 2.

A simplified overview of the client/server interactions and processes used in Maskiton. The client uploads/selects a stack on the server by selecting it in the web browser on the local machine. Once selected, the stack average is displayed in the browser, or if the averaging process is still running, as a series of updates until the full dataset is averaged. This progressive averaging occurs while incrementally splitting and converting the stack from IMAGIC to Xmipp format. The user can submit classification jobs without having to wait for averaging to finish, and the classification processes will reuse as much previously cached data as possible while returning their own incremental progress updates. The server caches a substantial amount of processing to the disk in the event that a future process can reuse a calculation to speed up the return of information to the client.