I was reading the article that Randall Kennedy, on InfoWorld, wrote about the relative ease with which Windows 7 installs, comparing it to each revision of Windows, going back to Windows NT.
He gives a picture of the problems of each of the operating systems, from his own perspective, and for the most part, I agree –
Windows NT: Setting up a Windows NT workstation was always a hit-or-miss proposition. Many common devices, like nonstandard IDE hard disk controllers, though perfectly happy under DOS/Windows 9x, would give Windows NT fits.
Any such reinstallation project usually required several days and involved at least one bricked attempt where a driver hosed the boot cycle and the Last Known Good Configuration option somehow lost it bearings. This was especially true when you added any number of unsupported third-party mobile power management tools into the mix.
Yes, there was always a point at which you knew the easiest fix was “Format C:” and then re-install. The bad part was over when you were done with the floppies.
Windows 2000: Like Windows NT, but with a half-baked plug-and-play mechanism to give you false hope that somehow the system would right itself. It rarely did, and most Windows 2000 installations quickly devolved into the kind of blue-screen troubleshooting sessions that were supposed to be a thing of the past. Still, a better-organized setup program — including some crude hardware autodetection — helped to shave a couple of hours off the marathon.
Actually, I really never had that many problems with the original install of Windows 2000. Perhaps it was because I never used it with 2nd or 3rd tier motherboards. I think that the BIOS was key in these cases, and the freaky, oddball AMI concoctions that some motherboards had was something to behold.
Windows XP: An attempt to meld the ease-of-use of Windows 9x with the robustness of the NT kernel, XP was a disaster out of the gate. Many Windows 9x drivers, though theoretically compatible with XP, simply didn’t work (NDIS driver support was a notable gray area). And some legacy NT drivers that worked fine under Windows 2000 simply refused to cooperate under XP.
It took nearly a year to shake out the whole hardware compatibility mess, and by then enough new standards and architectures had emerged to ensure that future XP installations would require much patching and tweaking (F6 for the driver floppy disk, anyone?).
I must lead a charmed life, because I have never had many problems with Windows XP installs. The hardest thing about XP installs was the waiting for that final reboot. It became less and less a problem as computers, and especially hard drives, sped up, getting down to full installs in about 12 minutes, though the installer always said it would be thirty-nine.
Windows Vista: Windows Vista had the potential to be the easiest version of Windows to install ever. And in most respects, it was — at least up to the point where you actually wanted to use it for something productive. Then it became a waiting game as you listened impatiently for news of Nvidia or ATI releasing yet another prerelease driver for your state-of-the-art video card, the one you paid extra for in anticipation of Vista’s Aero goodness.
And then it was install-and-pray time — pray that, maybe this go-round, the company had gotten it right. That you wouldn’t be rewarded for your patience and fortitude by a wonky release, with missing features or — worse still — the kind of instabilities that nearly sent you running back to XP with the last prerelease.
Eventually, the mess would get sorted out, with vendors accepting the fact that new technologies, like WDDM, were here to stay. But for a good year and a half after Vista went “gold,” creating a stable, working environment was a real challenge. No wonder so many of InfoWorld’s readers demanded that Microsoft do something and save Windows XP.
Indeed, it has been a long, rocky road that led me to this point — a point where I can install a version of Windows and have it successfully recognize the majority of the hardware in my system. A point where, even it if Windows doesn’t recognize something funky, Windows Update — that bastion of false hope in every previous iteration — will usually find a compatible (if generic) driver. And a point where a visit to my hardware manufacturer’s Web site rewards me with a healthy selection of new and updated drivers, all designed to ensure a smooth transition to Windows 7.
It was a long time coming, and perhaps the geeky side of me will miss the visceral challenge of a good old-fashioned, blood-pressure-raising, hair-thinning Windows installation. But the rational side is happy to see things finally improve to a point where Windows just works out of the box — and where the drama in my life is restricted to the day-to-day struggle of raising two disturbingly tech-savvy pre-teens.
Windows Vista was something I have only installed for myself twice, but I have re-installed it many times for customers. That was enough for me. Vista was, and is, a pain. But Vista is a pain completely due to Microsoft. They changed things that did not need to be changed, and kept trying to convince us that they were right in doing it.
I remain unconvinced, especially when I read what Mark Russinovich has said about the reasons for the changes.
In many of these cases, all of the blame was not on Microsoft, because many things were installed that were difficult to program. There used to be many types of odd hard drive controllers in the days of Windows NT, and today, there are not that many people who have a clue what they were, and are totally clueless when you talk about the many quirks that there used to be. Today, though not widely used, SCSI is a breeze to set up. When I started using it, that was not the case. I was kind of a SCSI guru, but it took many hours of frustration to get the title.
There were many other things that used serial interface and parallel interfaces, and the items did things that the designers of these ports never imagined. Since that was the case, sometimes there were amazing things happening when you tried to make the items work.
There were also things that happened that programmers of software were not prepared for, though they should have been. The rules of the game ( the IBM PC game, that is ) said you could do them, but instead of taking that into account, programs were designed stupidly (some still are) making assumptions about where they were installed, or should be installed. Virtually no one has 4 floppy drives on their machine any more (during the time from 286 to Pentium, I built every machine that was for my own use with 4 floppy drives, and they were sometimes A:,B:,C:, & D:, which meant that the first hard drive was E:); lots of programs did not like that.
I used to do things because you were supposed to be able to. I had several 486 class machines with 3 VESA Local Bus cards, which, though part of the standard, few motherboards were built well enough to allow it. It was not easy getting them to work, though it should have without problems.
Most people today look at the computer as an appliance. They are not as much a hobby pursuit as in times past. Also, with the loss of COM and Parallel ports, things are easier, with fewer surprises due to USB connectivity and standardization. The fact that hard drives are much larger means that fewer drives are seen on some enthusiast’s systems, and most likely they are all SATA.
Plug and play actually works most of the time, and there are much more standardized BIOS implementations. all of these things have contributed much more to the ease of Windows 7 than what Microsoft has done. It has all been laid at their collective feet. The fact that Microsoft no longer really worries about code optimization ( I’m talking the kind where it’s done by doing it by hand with assembler! ) also makes things easier.
So it’s been a long road, but Microsoft has not been the only commuter.
§
Quote of the day:
Writing gives you the illusion of control, and then you realize it’s just an illusion, that people are going to bring their own stuff into it.
– David Sedaris
⌘
⌘