CompuServe Thread

That's rather a dramatic heading, but I trust that after you've read the following, you'll feel its justified, especially as it supports the three major perspectives on this issue. That of Ana, David, and myself. There are other's who provided additional insight, but for me, these were the two that got me dealing with this issue. How does one reconcile all the differences of opinion on this subject without someone being wrong? Good question, and that't what I asked myself, because for a time, it was almost unbelieveable that we could be talking about the same OS. I was looking for a common thread that could explain the difficulties that David was having and the relatively smooth experience, I had had. At first it appeared that the way a system was configured was probably the likely difference. And in fact this could be a problem. So something to consider when you start getting into NT is that you know where all your IRQ's DMA's etc for each of your peripherals reside. However, the next issue was not so easy to explain, and that was of the multiple drives, that and the fact as also mentioned recently, that a large format hard drive had to be partioned in order to accomodate NT, really got me to wondering. Once you get into NT, I think you'll see, what I saw which is a very tightly organised systematic approach to an operating system. That being said, it was really difficult for me to imagine that NT was responsible for all these hardware dilemma's and problems. It just wasn't consistent with how its put together. Well I started looking through the README.WRI to try to determine the common element to each of the disparities involved with each particular piece of hardware. And that is where the answer was. Even though I'd read it several times, I didn't quite grasp the connection. The REAL truth is, that Microsoft had to make a decision about how to handle DOS and the harware that supports it. We know DOS is a quamire in alot of ways. So Microsoft was forced to make a decision on how to impose the loose standards of the DOS specificaton in a way that hardware issues could be resolved with the least amount of trouble, and yet provide a solid foundation for a sophisticated OS like NT. They did this by strictly adhereing to the DOS specification, figuring rightly I believe, that the vast majority of hardware would conform to those specifications. They must have known that this was taking the lesser of two evils approach, as we'll soon see. For instance, hardrives going back to DOS 3.3 which I believe is the version they chose to adhere to, could not exceed 1 gigabyte and also had a 1024 sectors limitation. Confronted with these limitations although not insurmountable 3rd party manufactures devised ways around them. This is where the whole issue of hardware compatibility begins. For instance Adaptec comes along and says well yes DOS has a limit of 1024 sectors but we can design a bios that will work around this so that we can offer controllers that will handle larger and larger drive capacities. Over the years many 3rd party applications have done similar things. The net result being that when NT was implmented and adhered strictly to DOS specification there was alot of hardware, which had circumvented those specifications. And that is the problem. Not so much in NT but in strict adherence to DOS, so that NT would be compatible will all the hardware available and at the same time allow for a much tighter operating system. In fact NT provides a means for resolving some but not all of these issues. Many of which have to be addressed by the OEM. Adaptec being a good example. As you can see this also explains the problems David had with 3rd and 4th IDE's. Fortuneately newer IDE's conform to DOS and NT with sectors in the 1024 range. But many older drives have sectors which exceed this value which because of modifications in the Bios of certain controllers were able to exceed what DOS could not otherwise handle. A problem recently raised with regard to using a SCSI as the boot drive, is typical of a problem with Adaptec controllers. One solution is to create two partions, as DOS spec's don't recognise space about 1 gigabtye. This is the NT solution to the problem, as it will manage both partions, formats etc. Another option is get an upgraded BIOS. If you check your bios and discover that its ver 3.08. or older it needs to be replaced, and having done so you will have full use of the drive as a single contiguous block of storage. In the case of NCR PCI SCSI the DOS limitation has already been dealt with to be DOS and NT compatible. A drive with multiple partions can be repartioned easily in NT, one of its other nice features. Another example is an older version EDSI controller like some that DTC has made which had bios to handle sectors in the 1600's. In those cases, the drive has to be reformatted to the original 1024 and increase to 63 sectors per track, in order to get the DOS portion of the drive close to its original capacity. The remaining capacity, NT is able to retrieve should you desire it. So the long and the short of it is, it is the limitations imposed by DOS that have caused the most problem as a result of manufactures finding ways around them. Now to get everything back to a uniform standard, hardware has to be compatible with I think DOS 3.3. I'm not sure but that seems to be the case. And this explanation, explains why I could have such a relatively easy time of it, because most of my equipment conforms to the DOS specifications in design and configuration, where David had a less easy time of it, because he had utilised extensions of DOS like a 3rd IDE etc. which forced him to comply with the original spec through additional software. In fact it is possible I could have had similar problems even with newer drives had I chosen to install more than 2. And it also explains Ana's perception that NT is plagued with hardware and incompatiblity issues. NT provides a means to adhere to DOS, and for the future a tighter standard so that problems like this won't be a problem in the futur, as NT has the built in flexibilty to expand, which DOS had but not in any structured sense. I hope that this brings into focus the problems people have encountered and explains the apparent discrepancies in the three perspectives on this issue. It may or may not make your life easier depending on what kind of hardware you have, it should give you some insight into where the problem can arise. And more often than not there are several solutions. And on that note, I personally plan to let the NT issue rest. I am quite satified that it accounts for each perspective fairly, and provides closure to this issue. It does at least address the main issue of concern over hardware compatibility. And in spite of the difficulties we encountered in expressing our views on some of this, I'd like to thank Ana for her heated exchange which inspired this effort, and David having the courage to speak up in spite of my original skepticism, and also for working through it so that I could determine what exactly was the real issue. In a way I was fortuneate to discover the commonanility of the problem, because without this complex issue having arisen, I like most people would probably not have discovered what was at the root of it, but just worked around it, to get a system functioning. So I really hope that this will provide the insight you need to get your systems up and operational as quickly as possible. Alot of feelings and energy has been spent to that end. Please make it count. 🙂 John