The results look promising. Would really be great to incorporate it into current live sites who suffer the permissions system overhead.
Since we are talking about permissions, it would actually be great if users are granted permissions on the "Add a Comment" feature. =)
Is there someone who volunteers to replicate Don's tests on a faster machine with larger RAM? and possibly on SCSI disk? I think the results Don posted are already significant enough, but it would be interesting to see this tested on various hardware.
