Amen! That’s one of the most obnoxious behaviors I’ve ever seen. Granted, there are system emergencies that could need to do that, but AOL’s finally getting a webpage I requested 3 minutes ago ain’t one of them.
For now, I’m taking a wait-and-see-but-be-cautious stance. Basically, I’m gonna try to get my friends to use firewalls when I get back to school, and offer Gibson’s opinion among others when people ask me about XP.
Congratulations. My brother’s Windows box crashed 3 times yesterday, under very little strain. We have now exchanged personal anecdotal evidence. I am not saying Windows itself is just sitting there and crashing – though that can happen, that is just pathetic; obviously an operating system shouldn’t die if it’s just sitting there. What I am talking about is other programs crashing and taking Windows with them, or Windows crashing otherwise as a result of running other programs. Make no mistake about it, this is Windows’ fault. On a better designed OS, a program crashing will never take the system down. Now maybe you are just lucky, or maybe you somehow know the uber-secret MS shortcut key combination that makes Windows stable as a rock. I don’t know. I do know that Windows has a reputation for being unstable. I do know that in my experience that holds true. I do know that a vast pool of anecdotal evidence from Windows users affirms my belief. I do know that even people within Microsoft do not try to maintain that 9x is a stable kernel: I know a person who is a software engineer for Windows XP and he has been telling me the 9x kernel is an unstable piece of crap for longer than I care to remember. (If you absolutely must hear it from him himself, I will ask him if I can give you his e-mail address; if he doesn’t want to give it out for whatever reason I will withdraw that statement.)
I can even give you some reasons for the instability: a faulty memory management model, DLL issues, everything running effectively as root, Windows’ bloated size, its overly integrated nature, its trying to be all things to all people, etc. By far the most serious faults are the first 3 I described, but the others hurt too. It’s a cumulative effect. Of those 3, the first, the memory model, is the worst problem. Allow me to quote Jeff Duntemann, from Assembly Language Step-by-Step, page 152 (in passing, this is an excellent book for someone looking to learn x86 assembly with no previous assembly experience):
As a final note, regarding your comments about scandisk and defrag: well, it can’t hurt, but it probably isn’t really helping. Neither scandisk nor defrag addresses any of the real Windows stability issues. Furthermore, even if they did somehow help, why should you be forced to run them? I don’t run equivalent programs in Linux. (Ok, slight lie, upon booting up I have what basically amounts to scandisk run – however, since I don’t reboot except to update my kernel, I effectively never run it.) You don’t run background programs because they might take down the system? That’s ridiculous. As I have stated over and over again, a program should never be able to knock out the OS. You (the generic you, though maybe the specific you too) are so used to doing these things that are – to an objective observer – just weird to keep your system running fine that you accept them as normal (rebooting constantly, scandisk, defrag, avoiding various programs, Windows reinstallation, etc.).
So tell me, SPOOFE, are you honestly trying to argue for Windows 9x stability as a general thing, or are you just stating that in your specific case you have not experienced the problems that are so common for other people?
I will call this a difference in OS design philosophy, although I’m not really sure that’s the true motivation. The Windows 9x kernel is big and bloated; the Linux kernel is sleek and trim, comparatively speaking. This is because the Linux kernel performs only the roles it should perform: it loads up, gets everything set up, handles very low level stuff, and does nothing more. As The Great Gazoo wrote in this thread, once an OS has been loaded, it should basically keep track of files and handle hardware. The folks at Microsoft would like for you to believe that all the Windows integration is essential, but all it really does is increase the size of the code (making for more bugs) and executables. In my view they do this to make sure that their products are on your machine, and that you’re using them, whether you want to or not. Feel free to not take such a cynical position if you want. To me the facts speak for themselves.
Exaclty. I personally LOVE XP’s operating environment, and have been telling people it’s benefits, but I’ve also informed them of the potential security concerns, and the drawbacks to running XP, and have provided input on how to better secure their systems (virus scan, firewall, don’t open unknown files, etc.).
I have no doubt that MS’s primary motivation is to provide the all-around user environment (They’re aiming to make money, and more features generally sell better).
Windows 9x isn’t unstable just because it’s big and bloated (It can be argued that XP is big and bloated, but it has been VERY stable for me), its unstability primarily stems from its 16/32 bit hybrid bastardization. While this may SOUND good in theory, when enacted it didn’t work out all that well.
But, you are right: We do have different OS design philosophy: In my opinion, a good OS provides all the features you may want to use on your system. My ideal OS would have EVERYTHING integrated to work together (Not saying this SHOULD happen, just kinda thinking that it would be nice if everything just worked perfectly no matter what).
Oops. I kinda phrased that line pretty poorly. What I mean to say is that MS is trying to provide enough services to make their product a more attractive choice to make more money, not necessarily to make it more convenient for consumers. The latter just happens as a result of the former.
No, the autocorrect anticipates and corrects the most common typos, like “teh” for “the”. If you’re trying to type “antidissestablishmentarianism” and accidently insert a “z” into the word, it won’t catch it.
SDP…
Just my specific case. Additionally, I’m stating that a lot (not necessarily most, just a lot, in my experiences) of “Windows crashes” are really the result of the crash of some other piece of software. Windows gets a lot of blame it doesn’t really deserve.
But don’t get me wrong… I’m amazed that Win98 has been behaving so well. The copy of Windows ME that my dad has on his Crapaq computer has crashed so often we’ve nicknamed it “Osprey”.
SPOOFE is quite correct. The vast majority of ‘Windows’ crashes are really caused by software over which MS has no control at all. Virtually all crashes of the OS itself are caused by bad drivers or (in rarer cases) even bad PCI implementations.
On the other hand, a better designed OS would be more tolerant of mistakes in applications. Win9x sadly traded an insecure memory and object model for much better backward compatibility with Win3x, which, in turn was designed around the memory limitations of the 286 processor.
Relatively speaking, I think Win95 was the most naturally stable of the bunch, with WinME being the worst, and Win98 being pretty good.
However, this varies greatly from machine to machine. SPOOFE If your Fathers PC crashes regularly, then there is undoubtedly some sort of configuration problem that could be fixed to solve the problem. Updating (or in some cases downdating ) the drivers for the sound card, video card and/or IDE controller will often completely solve a regular crashing problem of that sort.
My own machines (4 of them running either Win2k, or Win98SE) simply NEVER crash except when I’m testing new software or debugging. But then again, I’m quite careful about which drivers I install, and what software I allow to be installed in these machines.
Any software that does prove to be the source of crashes gets immediately removed from the machine. (Exception - I left Deus Ex on until I finished it even though it crashed my main machine occasionally. The game was just too cool).
Quicktime can also be a culprit on a Windows machine. The quicktime software on the PC is an amazing hack, I’m frankly amazed that it works at all. The Apple people are wizards on the MAC, but they SUCK at Windows programming. <sigh>
Probably, but since it’s HIS computer (he doesn’t like anything “changed”, even if it’s for the better), I don’t bother. Besides, it’s the Compaq OEM version, which is, IMOSHO, decidedly less stable than normal Windows ME.
Except the point that SDP was making is that an OS shouldn’t be vulnerable to third party software taking it down, and lots of OSes aren’t. Windows 2000, for example.
>> an OS shouldn’t be vulnerable to third party software taking it down, and lots of OSes aren’t. Windows 2000, for example
So get Win2000. Win9X is vulnerable because it is backward compatible which seems to be a more desirable feature. But if you do not need the compatibility you can just get another OS.
Oh, yeah. That’s just common sense, pure and simple. This sort of obnoxious behavior can’t be excused simply because UI design is a “new” and underdeveloped field.
The X Window system handles this problem correctly. MS Windows does not.
A well-designed operating system should let the user decide which window truly demands his or her attention. Unfortunately, instead of letting the user make this decision, MS Windows has a way of saying, “This is what you really want to be doing!” That same design philosophy can be seen in other Microsoft products as well. (MS Word, for example, which frequently moves the graphics around in a document – with no apparent consistency, and without first prompting the user for permission.)
Most likely, SPOOFE seldom installs new programs on his computer. The initial Windows installation is fairly stable and harmless; it’s only when you want to add more programs and (gasp) actually do work that makes it a-quiver. That’s part of the reason why many Windows-based IT departments don’t allow their users to install programs by themselves; it helps stability if they can only choose from a list of “authorized” (read: proven not to corrupt the system too much) titles instead.
Having actually taken a few courses in OS design back in college, I can state empathically that the idea of “integrating” a web browser, photo editor, or MP3 player into Windows is so much Microsoft bull-pucky. An OS is fundamentally a low-level traffic cop, managing hardware resources and coordinating the requests from the software applications above it.
But of course – there are all those software markets that Microsoft hasn’t taken over yet! Why let Kodak capture the home-photo-editing market when Microsoft can have it instead?
That’s in the works for Windows 2010: “It comes with every single program you’ll ever want! Never buy another software title ever again! Includes MS Office, MS Internet Explorer, MS PhotoStudio, MS Messenger, MS Solitaire, MS Quake, MS Putt-Putt’s ABC, MS Wolfgang’s Cookbook, and MS Newscaster, all standard! [small](Minimum system requirements: 2 GB memory, 600 GB hard drive, and a 50 GHz Pentium 7 processor)[/small]”
There were compatability issues, yes, but the companies did try to work around them by making their proprietary formats publically known. Nowadays, with the standardization of the Internet, the widespread use of common formats, and the rise of XML, this would be less of a problem – it is possible to have competition from multiple OSes without losing the ability to exchange data across platforms.
Not quite – more like, Microsoft dragged its feet on developing OS/2 (they were working with IBM, remember?), so OS/2 came out several months after Windows did. And also remember that Microsoft misled third-party developers about which OS they were supporting (Microsoft’s infamous “head-fake”, where they told everyone to write for OS/2 because they were backing it), so when Windows first came out, the only productivity programs that were readily available were Microsoft’s. See Wendy Goldman Rohm’s The Microsoft File for further details.
From a technical standpoint, OS/2 stomps Windows 3.1 into the dirt – it was far more stable, ran Windows applications, worked great in real-time situations, and handled multitasking and resources much better. But as Bill Gates knows very well, technological innovation means diddly-squat if you can sucker everyone into buying an inferior alternative instead…
::shrug:: I come across new programs that I think might be nifty several times a month, and I install them right away. Rippers, compressors, codecs, add-ons, mods, etc. etc.
And I resent the implication that I don’t do any work. Downloading porn is LOADS of work.
Actually, all commercial OS’s are vulnerable to 3rd party drivers. Only pure MACH implementations aren’t.
But Win9x is vulnerable to 3rd party apps as well. This is undesirable, but unfortunately necessary. It isn’t possible to write a secure OS for a 286. And, given the mass of software written for Win3x (which is designed for the 286 or better), there was a need for a 32 bit OS that would also run that software. That need dictated the design of Win95 to a degree that made it vulnerable to badly behaved applications as well.
OS/2 chose to set the line between compatibility and security in a different place, and thus failed. Incidently, prior to about 1992 (before OS/2 3.0) the core OS part of OS/2 was written by MS. IBM wrote the GUI part. It was only after the market chose Win3x over OS/2 that MS abandoned OS/2 and let IBM move it forward alone.
Until the MS had their bets hedged. They sold BOTH a robust OS and a compatible one. Once it was clear that the market had chosen the compatible one, they dropped OS/2 and focused all of their attention on Win3x and WinNT.
This isn’t speculation on my part. I used to teach people how to write device drivers for OS/2 back in the early 90’s when MS was pushing that as the next big thing. I was quite disappointed when it failed.
This is true of course. Technically speaking an OS does nothing more than control access to resources.
But this is also completely misleading, because it completely ignores the fact that application developers need more services than pure OS services in order to efficiently develop applications.
Saying that an OS is nothing but a traffic cop completely ignores the value of the new paradaigm that Apple brought to us. Today, the we use the term OS to refer to the entire operating environment that users work in and application developers depend on having available.
That includes, among other things, the Finder, the GUI shell and the presense of utilities like a calculator and a web browser. Technically these aren’t part of the OS, but you can’t sell an OS without them no-one wants to buy.
You sir, don’t know what the hell you are talking about. MS didn’t head fake, the were themselves surprised by the success of Windows. There were a bunch of divisions at MS (the compiler devision in particular) that were caught completely flat footed. Management believed that Windows was nothing more than a stopgap and testbed that would wean people from DOS to a GUI and from there to OS/2.
When Win3.0 turned out to be a phenominal success, Bill and his management team stopped, thought, and then turned the company and decided to go with success.
That turn was just as painful for people inside the company as it was for those outside. Even then, they kept on developing and supporting OS/2 for two more years until it was clear that the market for OS/2 just wasn’t going to come back. Only then did they drop it.
Nonsense. MS knew that OS/2 was superior, they frigging wrote the thing The only part that IBM developed prior to OS/2 3.0 was Presentation Manager.
I worked with Mach quite a bit and IIRC the drivers were in the same address domain as the microkernel, and thus could easily crash it. The “OS personality” was in a separate address space however.