ckp@grebyn.com (Checkpoint Technologies) (06/30/91)
In article <14317@pasteur.Berkeley.EDU> navas@cory.Berkeley.EDU writes: >In article <1991Jun27.170049.21231@grebyn.com> ckp@grebyn.com (Checkpoint Technologies) writes: >> Here's the thing: Game machines require only medium >>reliability. If a game machine crashes, you lose what? Your best high >>score? If your business computer crashes, you may lose untold $$ worth >>of data. So you guard it with things like, oh, parity RAM, so that >>you are assured that incorrect answers are not propogated. And you >>choose a vendor that has a reputation for reliability. > >In fact, the first thing our DB instructor said was that recovery was only >important for high end systems, on IBM-type systems you don't need to >worry about that. Interesting opinion. >Let me ask you, after a parity error hangs your system, and your DB system >has no recovery code, what are you going to do? Go to a backup. Nothing else to do. You did, or course, keep good backups. >So, you're saying that IBMs are better business machines because they have >parity RAM? You either have a system with FULL parity checking, or your >feeble attempts at covering your rear are likely to fail. I'm saying that businessMEN think IBMs are better business machines, in part because they have parity RAM. This is a marketing opinion. It's obvious that technical opinions differ. That's OK. >IMHO, of course :) Of course! -- Richard Krehbiel, private citizen ckp@grebyn.com (Who needs a fancy .signature?)