On clusters where a single login node serves multiple different clusters or at least multiple different simfactory ini files we traditionally put each ET source try in a subdirectory named by the cluster name.

This however is confusing to new users who will often log into the cluster then checkout and try to compile the ET in $HOME.

It would be nice to:

1. not rely on these subdirectories
2. get rid of sourcebasedir for everything other than the sync and login and --remote subcommands

‌

--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2423/avoid-using-subdirectories-of-home-on