Dvorak on How Microsoft Can Kill Linux 842
gewg_ writes "John C. Dvorak thinks he knows the way Redmond can kill Linux. Basing his premise on the relative dearth of device drivers available for Linux (compared to what is available for Windows), he sees an opportunity for the Borg to embrace and extinguish." From the article: "The immediate usefulness of Linux running under Windows is obvious. You can use all the Windows drivers for all the peripherals that don't run under Linux. Drivers have always been an issue with Linux as PC users have gotten spoiled with Windows driver support. Today's user wants to grab just about anything and not worry about installing it and making it work."
vmware (Score:5, Interesting)
What would MS-linux have I can't get from (Score:5, Interesting)
Plus, there are quite a few hardware devices that work in Linux and not all versions of Windows, for instance my Kensington SVGA webcam, fine in Linux, not available in Win2k.
Confused... (Score:2, Interesting)
They're using the wrong OS to begin with (Score:3, Interesting)
If they want to just install a device and go, then why are they bothering with Windows? Isn't that what Apple OS X is for?
Doubtful (Score:3, Interesting)
RTFA - nothing to see here... (Score:5, Interesting)
If MS developed an "MS Linux" as described, it would be one of many distributions. Even if it became "the dominant" one (the only good use for which would be to use the Windows drivers for devices Linux lacks driver support for), then stops supporting drivers for their own flavor of Linux... ummm... hmmm... what would happen? Oh -
Dvorak suggests that this somehow magically kills *all* of the different flavors of Linux. (Not *nix, he mentions only Linux).
He also alludes to some heretofore unknown, undiscovered-but-for-M$-lawyers hole in the GPL that would somehow allow M$ to pry Linux from the hands of the community into its control.
I RTFA'd twice, but John, you lost me on this. I can only guess you were looking for more hits to your column website from
Re:Don't click (Score:2, Interesting)
Number of times I've been forced to reinstall my entire windows partition due to a buggy driver scramming my whole box? dozens and dozens.
Number of times I've been forced to reinstall linux for the same problem: 0
All my peripherals (mp3 player, digital camera, printer, USB mouse / graphics pad, etc...) work on both windows and linux.
So, again, how are people spoiled by windows driver support?
TROLL WARNING - DO NOT CLICK TFA!!! (Score:2, Interesting)
Seriously, Dvorak has turned into a troll! Like, I mean, WTF? "Windows will run Linux as its secret sex slave and cut off all its oxygen. SCO was the trial run for that..."
Dvorak just made my permanent black list.
What about the opposite? (Score:2, Interesting)
When it crashes hard, kill it's pid and move on.
I hardly think it work well the other way around (as per FTA)
Now as for drivers.
I like ndiswrapper. Wish there were more like it. generic wrappers for windows drivers.
then the manufacturers that are pressed for cash can still "support" linux and the larger manus can develop native drivers (or release specs).
I know some will cry that given the choice, no one would bother with the native ones. Make the wrappers work well and it won't matter.
A little tale for you (Score:5, Interesting)
Windows XP took around 15 mins to install, with a couple of reboots. I then installed my nvidia drivers. Rebooted. I then installed my firewall. Rebooted. I then installed the drivers for the cisco aironet card. Rebooted. I then installed the drivers for my Delta-Audio 1010LT soundcard. Rebooted. I spent over an hour installing all the drivers I needed to make my system *functional*.
Suse took ~20-25 mins to install with all the software I wanted. When I logged in, everything just worked...
People say they use windows because it just works. Bull. It's just that people have been conditioned to accept that installing drivers is not part of the installation process.
There may be more drivers available for windows, but I'll stick with the linux way of doing things and buy cautiously.
Re:Confused... (Score:2, Interesting)
There will always be the Linux fanboys refusing to buy whatever Microsoft makes anyway.
scambled, as usual (Score:3, Interesting)
Funny how it only really works that way on Linux and OSX, in my experience. Dvorak's facts clash with reality, as usual...
How Linux can kill Windows... (Score:4, Interesting)
Re:Don't click (Score:4, Interesting)
If I plug in my Yepp MP3 player (A USB device), my video camera disappears (IEEE 1394).
I have a Soundblaster PCI 128. I have the XP drivers I DL from the website and the card doesn't work on any of the XP machines, as well as a 98 machine I plugged it into. But it works perfectly fine on my Fedora Linux machine, without any monkey business.
I consider myself a rational person, and to my rational thinking the driver support system for windows is on crack. (that, and/or Dvorak)
Re:They're using the wrong OS to begin with (Score:5, Interesting)
Under FC3 I plug it in, the computer automatically recognises an Ipod has been plugged in and makes the folder /media/iPod I type yum install gtkpod and I have a working iPod in under 2 minutes.
Over on Windows - and I'm not sure if this is the same on an Apple - it took about fifteen minutes of copying software from CD, signing up online, agreeing to several licenses, entering the serial number at least two times in different places...
I've had similar experiences with a Samsung laser printer that 'just worked' on linux but took an age to install on windows.
While I'll admit not everything is supported under Linux, of the stuff that is it seems a heck of a lot easier o get it running than with Windows.
Worked for NT 4.0 (Score:2, Interesting)
Of course the difference with Linux is it should be easier to integrate new drivers, once they are, um, written.
Re:Jesus, What a MORON! (Score:5, Interesting)
Today OSNews was talking about a Dvorak article in which he proclaims that if Microsoft created a Linux-Windows hybrid with the Windows driver layer, you would have instant compatibility with every device under the sun, and MS would take the Linux world by storm.
It's an interesting thought, and it brings up something I've been tossing around in my head for a while.
What if there existed an open standard for an operating system driver API? Such a standard would cover things like how the driver communicated with the kernel, how it was seen by the rest of the operating system, etc. If successful (and sufficiently free of restriction), it might be possible that many different operating systems would support it.
This would truly be one of the holy grails of computing: the device manufacturer would only need to write and maintain a single driver, and everything from Windows to BeOS would be able to use it.
Re:Confused... (Score:3, Interesting)
What makes Xandros buggy? Isn't it all the same basic sets of open source code that's so much more secure and stable than anything Microsoft produces?
Responsibility of the manufacturer or OS? (Score:3, Interesting)
On the one hand, you have the market force that would make HW manufacturers want to provide quality drivers to the three OSes. On the other, you have the OS companies that want to support many drivers.
3 Problems (Score:4, Interesting)
1) Windows Drivers Suck. They are often buggy, and bring the system down. I don't want my Linux system dependant on buggy Windows drivers. I'm happy with my linux system as is (yes, you have to do some research to make sure what you buy is compatible. That's life--- Be an educated consumer).
2) Inane amount of difficult involved. The Windows driver model is VERY different from the Linux driver model. I'm not a sure a 'hybrid' is possible without a great deal of work/new code. Do you really want a Linux where MS wrote 1/3 of the Kernel?
Especially if that portion is closed source? Who knows what bugs/exploits will lurk there. No Thanks!
3) The Linux driver model is superior. I can take my harddisk out of my desktop (with ACPI on), and drop it into a desktop with a different processor, different network cards, different motherboard chipset (with ACPI off), different graphics card, and it'll boot. On SuSE, SaX2 will run automagically, press enter a couple times, and *Poof* you're up and running.
Try this on Windows. Blue Screen, almost certainly.
Does the Windows Driver Model permit dynamically loaded drivers? I think not.
Does the Windows Driver Model require a reboot on each driver installation/upgrade? Depends on the device, but usually.
Does the Windows Driver Model support having thousands of drivers installed simultaneously, and dynamically loading the necessary ones on demand?
I think not.
No thank you. MS-Linux will only draw people from Windows, not Linux.
Once you go to the pain of making sure ALL your hardware is Linux compatible (i.e. working drivers are out there), the Linux driver model is preferable to the clunky windows driver model.
Yes, I know there are reasons the Windows driver model is the way it is. Mainly backwards compatability. But rational != excuse.
Linux is better, and I like it that way.
Re:They're using the wrong OS to begin with (Score:3, Interesting)
Okay, so it was only two things, why bother with the list. Methinks your 15 minute adventure was due to either your slow typing or a slow CD-Rom drive. Did you count the time it took you to figure out that 'gtkpod' was the correct/best software to install under Fedora?
I like OSS as much as the next guy (unless the next guy is RMS), but the incorrect Windows to Linux comparison is soooo 1999.
spoiled by Linux, instead. (Score:2, Interesting)
There is a lot of hardware that doesn't yet work with Linux (which I wish did); for many people, this is a real downside to Linux as an OS choice right now for two closely linked reasons: 1) it means their existing hardware might not work with Linux, which might mean spending money to replace them if Linux is for other reasons especially attractive 2) When a new device comes out, especially a specialized one they might need in a field like medical imagery, it's likely to come with Windows drivers (if it's designed to interface with an PC, rather than self-contained) but may never work, or may only work with reduced functionality, with Linux.
However, there's another side to the driver problem. Adversity breeds strength; those devices which do work with my installation of Mepis Linux (and which have worked with various other distributions) generally don't need separate driver downloads to make work; my printer, for instance (Lexmark 210e) works under KDE after a 2-minute exercise with an actually decent "wizard" type application. (I loathe those "Wizards" in general, but this one works nicely, isn't condescending, and results in a working printer.)
For reasons unimportant here, I recently had to use a machine running Windows 2000; to make it work with the same printer, I had to find the driver for the printer and install it. (Which, surprising to me, did not require a re-boot. Thanks, that was less unpleasant than I expected.) Likewise, and more annoying, the same was true of an ethernet card I hooked to the same laptop. It came with a driver on floppy; none of my machines have a floppy drive. Luckily, I had a USB thumb drive handy, could download the driver from a different machine, transfer via the thumb drive. The same card is auto-recognized under Linux and Just Works. Perhaps it's also easy traveling with WIndows XP, but I don't have that to compare.
My dad has a color laser printer (Minolta/QMS) which for about half a year would cause his Windows-running computer to get even crummier whenever it was used; Minolta tech support blamed a memory leak in the driver or the spooler software. I think a new driver has solved the problem, but in the free operating system world. the problem *might* have been solved a lot faster; in the time between discovering the gooey performance under Windows and an improved driver, that model of printer actually gained support under CUPS.
I also have some older hardware for which drivers exist for Windows 98 -maybe also 2000-, but not for XP; that means that for many users it would be effectively useless. (Or do those drivers also work in some sort of compatibility mode for XP? Haven't tried, don't know.) Most of it works fine under Linux.
So while it's nice for the buyers of new Windows-centric peripherals that they can install software to make the peripherals work, it's also nice not to be dependent on separate driver software. (Which, Yes, is needed to make some things work under Linux, too -- there's an overlap, clearly.) And for ethernet cards, it's plain annoying.
So while using Linux as my every-day desktop (as I have for the last 6 years) limits my hardware choices, it also means that my printer, scanner, modem, etc. really *are* plug-and-play -- or at least closer than I've ever seen to that ideal under Windows. [Note, YMMV; I find Windows annoying enough that I don't ever deal with it at much of a stretch. The laptop in question will soon be upgraded to a nice Linux install
timothy
Seems unlikely (Score:3, Interesting)
GPL would be quite a problem for MS here. Binary driivers are OK IFF they use nothing but a subset of the functions exported to modules. Other useful functions are exported only to modules that declare themselves to be GPL. No promises are made that any particular function will or will not remain available to non-GPL drivers.
The more interesting Windows drivers would be the third party ones for brand new hardware. MS doesn't own those, so if they want them to work in Linux, they'll have to come up with a full translation layer under the GPL. Native GPL drivers (by avoiding extra layers of bogosity and being open for improvement) will always be superior under Linux. For that matter, they tend to be superior to the Windows driver under Windows.
As for availability, I find that a recent Linux kernel is MORE likely than Windows to come with the needed driver. This is especially important for network drivers where you can't just go download it if you need it. The last example I saw was for the BCM5701 network controller. That is a fairly common builtin Gig ethernet chip (especially for AMD chipsets) that Linux has supported out of the box for quite a while.
No said yet (Score:3, Interesting)
Windows NT, by most accounts, was a solid OS design, partitioning and securing different parts of the system from on another. But it was fast enough to beat the competition on every benchmark, so Microsoft made the fatal decision to move the video driver into the protected kernel space. Thereby, damaging the OS stability.
They would expect to do the same to Linux. Play games and take shortcuts with the system stability, so that a fault in one system would bring the whole computer down. If they ever did try such a monstrosity, I think most user would totally reject such flotsam.
Dvorak has absolutely NO credibility (Score:3, Interesting)
I'm not trolling here. It's a serious question. The guy is the quintessential know-nothing tech writer who seems to have figured out how to thrive by writing utter hogwash.
Seriously. Name one thing in the last five years he's actually gotten right.
Re:WHAT!? (Score:5, Interesting)
The Macintosh uses an experimental pointing device called a "mouse". There is no evidence that people want to use these things.
- John C. Dvorak, SF Examiner, Feb. 1984.
Re:Jesus, What a MORON! (Score:3, Interesting)
This is why for example Sun had/has a hard time including its drivers in open solaris.
A GPL'd program can run on a proprietary OS. No question about it.
But if you want to take a GPL OS like Linux and add proprietary drivers and release it as a product, you're in violation of the GPL and you automatically lose the rights granted in the GPL, meaning, that Microsoft would not have the right to use Linux in their software at all.
However, if they did release the drivers in the GPL, then the next hacker with broadband could say: Gimme all the source code (a requirement of the GPL) and make install, make, cut
They can't. Because they can't. Not only would they make absolutely no money from it, but they would also lose all the money they woud have made selling binary windows software.
They've screwed themselves in a corner, and they know it. They can't deliver longhorn. (I mean a stable operating system, secure, fast, and able to compete on price/features with linux). They won't be able to. In december 2005, it will be painfully obvious that nobody's lining up to buy it, and they'll not release it because it would cost too much to market it. Rememebr that by that time, Gnome and KDE will have iterated throught another enhencement cycle and that suse and redhat will have deployed desktops enterprise-wide at large companies.
Sun will be fried already, and either MS will buy it to get Java and J2EE, or it will collapse in bankruptcy court and sun.com will be taken by a domain-monger.
Microsoft is very bad at adapting to business conditions. They have a high and mighty attitude like "We know what people want and we're going to make sure they get it". Firefox has completely disproved that theory by making the browser people want, because MS wasn't.
MS is flying high at 80,000 feet, thinking that their market valuation and excellent products will enable them to not only ride the wave but win. I say the opposite is true. Their cash is a liability. It's not their money. It belongs to the shareholders. The shareholders either want the money put to good use, or returned to them so they can put it elsewhere (like wheat futures or whatever). Their products are outdated, riddled with fundamental security flaws, and too expensive.
Their cash cows are Windows and Office. Both are tenatiously riding on the need by companies and individuals to maintain business continuity with existing applications and documents.
All new application development at companies are browser-based or OS-agnostic (python, java, ruby) , and one can do just fine with knoppix for browser-only.
Word and powerpoint are going to OO.org. Excel spreadsheets on analyst desktops are going into more robust and extensible database back ends. The custom access and vb applications are being phased out in favor of web apps, and finally, old office documents will be zipped and burned to CD, only to be looked at when auditors show up).
Then, companies will have no need for windows and office. Do you think home users can support 2 billion dollars a month in revenue? Nah, it's over.