If you launch a notebook using dataset A, leave it up, then try to launch another notebook using dataset B the browser just loads up dataset A's notebook.
If we made it so the hub didn't store the user id in the local browser cache then each time a browser is opened to a notebook (even if the same notebook) they'll get a new instance in ec2, new user home dir, etc. Is this what we want?
If you launch a notebook using dataset A, leave it up, then try to launch another notebook using dataset B the browser just loads up dataset A's notebook.
If we made it so the hub didn't store the user id in the local browser cache then each time a browser is opened to a notebook (even if the same notebook) they'll get a new instance in ec2, new user home dir, etc. Is this what we want?