Currently, it seems that when the browser crashes or the browser window is closed, the experiment stops and all settings (like OD calibration, pump rates etc.) revert to their defaults. It is this by design or a bug/oversight? It would make long-running experiment less prone to failure if one could simpyl re-connect to an already running experiment from a new browser window.