ruby - when to use flush after writing to STDOUT? -

i know flush does. takes data stored in buffer , writes standard output empty buffer. not clear me when needs done (shouldn't flushing done automatically interpreter). reading book ruby programming language , gives example of socket server:

require 'socket'  host, port = argv  begin       stdout.print "connnecting..."      stdout.flush      s = tcpsocket.open(host,port)      stdout.puts "done"       local, peer = s.addr, s.peeraddr      stdout.print "connected #{peer[2]}:#{peer[1]}"      stdout.print "using local port #{local[1]}"       begin          sleep(0.5)          msg = s.read_nonblock(4096)          stdout.puts msg.chop      rescue systemcallerror      end       loop          stdout.print '> '          stdout.flush          local = stdin.gets          break if !local           s.puts(local)          s.flush           response = s.readpartial(4096)          puts(response.chop)      end  rescue      puts $!  ensure      s.close if s  end 

why invoke flush after writing "connecting" stdout, not invoke flush after writing "done" stdout? seems arbitrary here. confusion is.

to ensure "connecting..." gets printed out before potentially long-running thing happens.

once you're connected it's less important immediate feedback, , you're flushing after ">" characters printed in response reading data.