gus@Shasta.ARPA (Gus Fernandez) (02/23/86)
> > During that time there have been two drives replaced out of 4 IBM PC's > > at work. Those are XT's so we are talking about a 50% failure rate. > > Ok, I'll bite. The point is, at least on a IBM, when the drives fail, the > DRIVES fail, and quit writing disks. This is because the MFM controller > uses a reasonable format, and can detect bad reads and crash the machine > immediately. The most common failure mode we've seen at Reed on the 400K > drives (I would estimate half of our 500 odd machines have this problem > to some degree) is misalignment just severe enough to cause OCCASIONAL > miswrites, which make the disk readable ONLY BY THE MACHINE THAT WROTE IT. > And of course there's no way to realign the drive. All Apple will do is > replace it. Neat, huh. > The controller has nothing to do with it. The problems you are having are strictly mechanical hardware related. The IWM puts out just as reliable a disk format as anything that an IBM machine spits out. The IWM is also no slower than other disk formats. The reason that Mac disk access seems slow is that the Mac uses a much more sophisticated two-layer file architecture that IBM. By two layers, I mean first the file system layer, and then the resource layer. Generally such systems are slow at first in that common inter-layer interactions are not optimized. The new ROM's solve this problem in at least FOUR ways; 1: Track cache 2: File cahe 3: Resource manager optimizations 4: Supercharged resources. I am not even counting HFS and 800K drives. The problem with error reporting is indeed a sticky one. Should a floppy based OS display an MS DOS read error message a la MS DOS or should it handle errors internally. There are arguments both ways. One argument says that in the case of a read error, the user should be informed that something is wrong. On the other hand, by then it may be too late anyway. Case and point. I bought a copy of EASY3d last month. The supplied disk had a manufacturing flaw (a read error on one of the sectors.) The program worked fine but I kept on hearing the disk recalibrate every time I launched the program. The program could also not be copied even though the label said that it was not copy protected. I want in with the resource manager and a sector editor and found that indeed there was a bad block, but I was able to recovering using bit-level microsurgery. The result, a new menu item appeared on the screen that was not there before. Now try THAT with an IBM! > At any rate, 50% PC floppy drive failure rates are unusual, in my experience. > Besides, many PCs contain third party floppy drives. Try that on a Mac! Several third pary floppy drives ARE available for the Mac. The habba drive and the maple drive are two examples. > Admittedly, not all Mac drives are bad. But the percentage rate, here, > is high enough to make me angry. Especially when my machine at home is > one of them. > -- Nobody should jump to conclusions with a limited sample, even if its there own. The aggravated shout the loudest. The vast majority are silent. > Bart Massey > ..tektronix!reed!bart Gus Fernandez gus@shasta