> Quite apart from that, it's a *dumb* idea.

> The kernel time would be about 30 seconds out form wall time,

This statement doesn't even make sense, never mind being true or false.

Kernel time is an integer, the number of seconds since somewhen
(actually, microseconds, a distinction we can ignore for the moment).
This is not how wall time works anywhere, at least not anywhere *I've*
heard of.  Thus, a "difference" between kernel time and wall time
simply does not make any sense.

If you mean "kernel time, converted to a wall-clock time with the
current timezone info files, would be about 30 seconds out from wall
time", that may well be true, but what it actually means is that to do
this right, you have to adjust the timezone files to deal with leap
seconds when converting between kernel time and wall-clock time.  (And
maintaining them as leap seconds are inserted could, as you point out,
get interesting, and that _is_ a problem with the suggestion.  A
relatively minor one for most people, I daresay, but still a valid
problem.)

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B