[pmoriarty]

Your point is taken, but in this particular example, you should just use the right tool for the job. In this case, the "stat" utility. For example:

  > stat --format='%z' *
  2016-05-27 12:39:46.559137232 -0300
  2015-07-26 14:37:51.714193856 -0300
  2016-04-15 19:55:33.329346654 -0300
  2016-03-21 02:59:05.377041620 -0300
  2015-11-22 19:27:56.868541801 -0300

Sorting them and picking the earliest would just be:

  > stat --format='%z' * | sort -n | head -n 1
  2015-07-26 14:37:51.714193856 -0700