HOME
TOPICS
SEARCH
ABOUT ME
M AIL

 
 		  technofile
Al Fasoldt's reviews and commentaries, continuously available online since 1983

What should you believe when someone says 'Windows never crashes'?


Jan. 1, 2001

By Al Fasoldt
Copyright ©2001 Al Fasoldt

   Every now and then someone tells me something like this: "I've used Windows for years and have never had a crash. Why do you keep harping on Windows? It's not as bad as you say."
   What am I supposed to make of this? Is it possible that some Windows users never have a crash? Let's suppose we are all experts on poisons. Would we believe someone who swore to us that a guy in Topeka has been swallowing arsenic for years without getting sick or dying? Surely not. And that is precisely how I feel about these assertions about miraculous stability in Windows 95, Windows 98 and Windows Me, the three consumer versions of Microsoft's operating system. No way. No way at all.
   Yet we can imagine scenarios in which Windows would behave better than in normally does. We can consider them in a moment. There's another possibility I'd like to dispose of first. Perhaps the Windows users who assert that their PCs never crash just don't know what a crash is. As remote as that might seem, this possibility is attractive for two reasons.
   First, although we should be able to agree that a non-responsive (or "dead") Windows PC is outright evidence of a crash, users who are not technically inclined might not realize the difference between a crashed PC and one that seems to be waiting for something to happen; both, after all, are non-responsive. After all, if your Windows computer refuses to respond for a few seconds and you decide to reboot it, has Windows actually crashed? Or did you run out of patience? I'm not suggesting which answer is correct, but I do believe we need to be less dogmatic about this issue.
   Second, some Windows users might not realize that an operating system has full responsibility for what happens within the scope of all software running on a computer. A properly designed operating system probably can't prevent individual programs from crashing, but it must prevent one program's actions from interfering with the operating system itself and with other programs. In other words, an operating system that is designed right will make sure that Program A will keep running even if Program B crashes.
   The consumer versions of Windows don't do this, as most of us already know. But implicit in my statement is a key that could unlock this mystery. Someone unaware of the technical role of an operating system might assume that a Windows crash caused by Program B is not a Windows crash at all. The reasoning behind such an assumption might go like this: "Program B wasn't designed right, and that's why everything crashed -- and you can't blame Microsoft if somebody else makes defective software."
   Microsoft itself encourages this kind of relativism. Windows is famous for reporting that "Program X performed an illegal operation and will be shut down." No one should be blamed for assuming that such apparently clear phraseology hides a meaning that reveals itself only over time: "Windows was not able to prevent Program X from performing an illegal operation. Windows is now asking Program X to shut down, and if things aren't better soon, Windows will crash."
   Look carefully at the two phrases. The first transmits a message written by lawyers -- "A person entered your home while you were sleeping and left through a window while carrying your jewelry box" -- while the second tells it like it is: "You forgot to fix your home security system and the police were nowhere to be seen, so a crook jimmied your window and stole your precious jewelry." The first message assays no blame; the second says it like it is. Microsoft crafted the "Program X performed an illegal operation and will be shut down" message in an attempt to keep users from knowing what an operating system is supposed to be doing. If Microsoft were in charge of transporting school children, bus drivers would never be responsible for what happens on the bus, and reports of misbehavior on the bus would say merely that "Johnny R. performed an improper act and was ejected from the bus." You would never know what the "improper act" was, nor would you be told that for weeks the bus driver had been ignoring pleas from all the other children that he make Johnny behave.
   An operating system that does not operate properly is a serious problem. But perhaps more serious is the damage Microsoft's "Newspeak" has done to the public's understanding of basic responsibility. Windows users who say they have never seen a crash could be speaking that new language, as absurd as it might seem. (In case you believe I am exaggerating, consider the message Windows displays when you boot up your PC after a crash. It's a simple matter for Windows to know that it crashed the last time it ran, yet the message says someone else (or something else, an ominous shading of the meaning) caused a problem: The operating system is forced to scan for errors "because Windows was not shut down properly." Not, mind you, "because Windows crashed" or "because Windows was unable to prevent a crash." Double-talk is double-talk, Orwellian or not.
   But let us suppose that there are, in fact, Windows users who are honestly reporting that they have never experienced a Windows crash. What could make that possible, and how could we learn from it?
   I cannot imagine more than three scenarios:
    Windows is not used to run any programs. The computer sits all day without doing anything more complicated than popping up a calculator.
    Windows runs only the SAME program all the time it is running. The operating system might never develop conflicts.
    Windows is rebooted often. I know some users who reboot every half-hour or so. When I ask them why, they usually answer in one of two ways -- that they like to make sure Windows starts out fresh every time it does something different, or that they can sense that Windows is slowing down and needs rebooting.
   Until I started running Linux two years ago, I, too, rebooted Windows often, and I did it without a second thought. It did not occur to me that computers should not need to be rebooted; I simply assumed that rebooting was a normal part of each hour's routine. (A personal note: My biggest surprise when switching to Linux for all my daily work was not that my PC no longer needed rebooting. It was that my PC behaved perfectly for the first time. Apparently I had been blaming the PC's hardware for the frailties of Windows.)