At 07:25 AM 8/31/03 -0400, David H. Bailey wrote: >But I do need to say that I >have never (knock on wood) had an ATA drive "crumble" at all, let alone >every 10-12 months.
At 08:03 AM 8/31/03 -0400, David H. Bailey wrote: >I remember a doom-and-gloom message from a computer-industry worker >concerning commercial CDs, since nobody really knew how long the >substrate would last, but I have never had one of those fail, either and >the ones I first bought 15 years ago still play just fine. Anectodal evidence of success only says that you haven't seen a failure yet. When talking about reliability, the incidence and cause of failure are much more interesting. Neither is scientific, but I've learned lots more about tending my hardware garden from failures (mine and others') than from anyone saying "it works just fine". (Then there's the "it works just fine until..." and the "it works just fine, except for the occasional blue screen of death..." and the "it works just fine, except for the occasional bomb...", but those are other stories.) I have had CD rot with commercial CDs. Some, like radio station promos, are expected to break down because the reflective layer isn't sandwiched, it's just below the silkscreened label. Others, like the new Disney DVDs, are intended to provide only a few days' use and then self-destruct. But I have in my collection several CDs (about 5 that I know of out of more than 3,000, a small ratio) whose improperly sealed edges have let in air, causing the aluminum reflective surface to oxidize, and whose cracked centers have done the same. I also have one Telarc CD that looks fine, but just won't play anymore. (That's okay; it's only Beethoven's 7th.) I've pointed out (again anecdotally) five hard drive failures on my own machines. That is over a span of 10 years with a total of 22 drives. I would still consider that a high failure rate. So why so high for me? What would account for this? There are lots of reasons for drive failures, but I would suggest that mine are due to heavy use in multitrack audio. Such multitrack audio is often writing or reading 20-40 files in real time -- lots of arm motion, lots of stress on delicate solenoids, and lots of heat buildup around drive transistors. These are nearly equivalent to manufacturing stress tests, but in real life. And as newer drives are 'on their own', so to speak, using bus mastering, they run at the highest possible data transfer rates -- and hence get clobbered by demanding users of multitrack audio who won't give it up until they get audio dropout! (Those interested in drive testing can do web search for Jose Catena's DSKBENCH, or I can send it along. I will email it if you can't find it. It's a 41K program that runs in a DOS window.) There's another issue that's victimized me once: sticky drives. My machines are in use 24/7, processing audio or video when I'm asleep. The only time they are turned off is when I am away for more than a day. I have learned to back up hard drives before leaving because drives in constant use have some effect on their bearing lubricants that I don't quite understand, but has become known. They become sticky when they cool, the platters freeze and won't spin up, and the drive transistors blow. $1,000 data recovery charges in the future! (I won't describe how I unstuck that drive; it's a don't-do-this-at-home trick involving kitchen stoves.) Our list appears to be a relatively conservative group of hardware users (at least we haven't heard much from those who upgrade hardware regularly), whereas multimedia folks and gamers are fierce about upgrading hard drives -- and gamers are a huge market. Just look at all those third-party video cards on your average office supply store shelves. Those ain't for middle managers! This started out with Philip Aker's comments about ATA vs. SCSI reliability. I guess, to summarize -- and apologies for so much verbiage, but I hope some of it was helpful -- I feel it's wise to upgrade regularly not only because it gives you more room for expansion but also because it 'resets' the likely failure time to zero. To use the cliché, "it works for me". :) (Yeah, I know, more useless anecdotal evidence.) Dennis _______________________________________________ Finale mailing list [EMAIL PROTECTED] http://lists.shsu.edu/mailman/listinfo/finale
