This would allow running the backend on some persistent server and then returning to the calculation anytime, anywhere. This would also allow for testing or doing background calculations without a browser, which may be convenient.
This would mean the computation history would move back to the backend instead of the frontend, but we we can deal with that.
The backend should be capable of simultaneously running several unrelated computations (e.g. if it has more than one user).
The only issue to think about is to make sure it is not easy to accumulate "abandoned" calculations, because that can be pretty taxing on memory.
This would allow running the backend on some persistent server and then returning to the calculation anytime, anywhere. This would also allow for testing or doing background calculations without a browser, which may be convenient.
This would mean the computation history would move back to the backend instead of the frontend, but we we can deal with that.
The backend should be capable of simultaneously running several unrelated computations (e.g. if it has more than one user).
The only issue to think about is to make sure it is not easy to accumulate "abandoned" calculations, because that can be pretty taxing on memory.