It is alpha 1 going to alpha 2
I doubt we can expect anything better. You can blame serious bugs for the gap between alphas. And without recoverable data, then every time the alpha 1 network died and restarted the network would come back up and app developers & users have to restart loading data/accounts again. Maidsafe understandably decided against restarting and using up their resources to keep restarting a bugged alpha 1 and using their resources elsewhere.
Also mock crust/routing can help as a stop-gap
Good news is that as we go through the alpha stages this should not happen as much. And when data republish is implemented the data won't be lost either. One cannot expect a network to be constantly running until beta (my opinion/experience)