Perhaps the problem is that it was, in fact, adaptable, meaning that if all the supernodes your client knows about go down you're no longer able to bootstrap.

Perhaps the problem is that it was, in fact, adaptable, meaning that if all the supernodes your client knows about go down you're no longer able to bootstrap. In fact, I found I was able to reconnect faster after the outage by deleting the shared.xml file that stores known supernodes - presumably connecting back to Skype-operated bootstrap servers.