On Wed, Jul 23, 2008 at 06:47:04PM -0400, J. Bruce Fields wrote: > Right. My impression was that they are real first-class inodes that > happen to be looked up in a strange way. They are - at least in the Solaris UFS implementation.