Yes, I was thinking of how to run RightEdge optimization/simulation on multiple machines automatically. After some consideration of the above two ways, I might prefer the first way in general, because it allows optimization tasks to be pushed to (first way) rather than requested by workers (second way). The former sounds less complex and more extensible.
But I have some doubts with command line in the first approach, because it doesn't seem like the command line can do all that the UI can do. For example, the command line can perform historical data update, but it can't do auto fix according to the command line documentation in case the data updated have errors. Another is how to sync folder definitions across machines.
I don't know if extending command line functions would be a massive change to RightEdge, but another way seems to be able to run RightEdge as libraries in a custom process, so that the sort of client server communication things may be handled by the custom process rather than bundled in the plugins, which may be better to focus on its own tasks solely, for example, optimization plugin, while still maintaining RightEdge's flexibility.
Could you please have some comments on this?
Sunday July 07 2013 by