Hi, Keith.

KR> RS> machine isnt in use when you are at the other machine etc. With
KR> RS> an intelligent cycling of the actual tapes. Presumably with a

KR>  FM> "Intelligent cycling" - that's the question I'm
trying to get to! What
KR>  FM> is an "intelligent" cycle which addresses each of
those different
KR> needs
KR>  FM> above, plus any I haven't thought of?

KR> use the it industry method. a full backup, followed by a bunch of
KR> incrementals. keep on with incrementals until either you fill a tape or
KR> get to the point where applying all the incrementals to the most recent
KR> full backup becomes too much of a job, then start a new cycle with a
KR> full backup. you get much more generations for a buck that way.

I'm well aware of that and I think it addresses the real questions in a
very haphazard way. What I really want is what we created in the shop
where I once worked, *every* file has a specified number of generations
and/or time which will be retained, *and* a number of separate
copies of each generation, with useful defaults. I actually wrote some
of the s/w to do it, and the whole thing was intimately bound to the
o/s.

The approach you advocate still runs the risk of multiple generations of
perfectly OK files killing the one you stuffed up on only 1 generation
ago.

KR> RS> Still, the tax man pays for a substantial chunk of it.

KR>  FM> Some of it.

KR> come on frank, should be 47% (:

Some of us don't do it like that. :-)

Regards, FIM.

 * * What I used to do all night, now takes me all night to do
@EOT:

---