> The wiser maintainers can't generalize a 'ulimit' solution, yet > expect ordinary users to figure out specific ones. How?
Seconded! > It's about finding practical formulas for two variables: the proposed > job load, and the system speed. It's virtually truck driving**, > complete with freight limits, speed limits, time limits, mechanical > failure, spoilage, traffic, and driver fatigue.... > > [demonstration elided] In my opinion this is the right way to tackle this problem. Like John Roebling I would then add a factor of ten to cover what I don't understand. Kapil, if this kind of thinking is not already familiar to you, hunt up Jon Bentley's old 'Programming Pearls' columns on 'the back of the envelope' and (if memory serves) 'the return of the envelope'. http://cs.bell-labs.com/cm/cs/pearls/bote.html NR -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
