Posted on 07/23/2008 5:54:47 AM PDT by twntaipan
When you buy a new PC today, unless you hunt down a Linux system or you buy a Mac, you're pretty much stuck with Vista. Sad, but true.
So, when I had to get a new PC in a hurry, after one of my PCs went to the big bit-ranch in the sky with a fried motherboard, the one I bought, a Dell Inspiron 530S from my local Best Buy came pre-infected with Vista Home Premium. Big deal. It took me less than an hour to install Linux Mint 5 Elyssa R1 on it.
As expected, everything on this 2.4GHz Intel Core2 Duo Processor E4600-powered PC ran perfectly with Mint. But, then it struck me, everyone is talking about having to buy Vista systems and then 'downgrading' them to XP Pro, how hard really is it to do that.? Since I had left half the 500BG SATA hard drive unpartitioned, I decided to install XP SP3 on it to see how much, if any, trouble I'd run into. The answer: a lot.
First, thanks to my Microsoft TechNet membership I could download an XP disk image, which included all the patches up to and including SP3. Many people aren't going to be that lucky. They'll need to install XP and then download perhaps hundreds of megabytes of patches. Boy, doesn't that sound like a lot of fun?
If you don't have a MSDN (Microsoft Developers Network) or TechNet membership, there are two ways to approach this problem. The first is to manually slipstream the patches into an XP installation CD. You can find a good set of instructions on how to do this in Slipstreaming Windows XP Service Pack 2 and Create Bootable CD. While the article is for SP2, the same technique works for XP SP3 as well.
The other way is use nLite. This is a program that allows you to customize Windows XP and 2000. While it's primarily so that you can set up Windows without components you don't want, such as Internet Explorer 6, Outlook Express, MSN Explorer, or Messenger, you can also use it to create fully patched-up boot/installation CDs. I highly recommend it.
This time I didn't need to use either one. I simply put in my newly burned XP SP3 CD and went through the usual XP installation routine. Within an hour, I was booting XP.
If this had been Linux my work would have been done. With XP, I soon discovered my job was just beginning. I soon found that XP couldn't recognize my graphics sub-system, a totally ordinary Integrated Intel GMA (Graphics Media Accelerator) 3100; the audio system, the Realtek HD Audio chipset, or, most annoying of all, the Intel 10/100Mbps Ethernet port. How can an operating system in 2008 not recognize an Ethernet port?
Well, XP doesn't.
Fortunately, Dell includes a CD with the full range of Windows drivers on it. With it, I was able to install the drivers for all the equipment without much trouble. Within another hour, I finally had a working XP SP3 system.
That wasn't so bad was it? Well, here's my problem, except for Dell, I don't know of any vendors who ship their PCs with driver disks anymore. The usual vendor answer for when you have a driver problem is for you to go online, search down the right driver, download, and install it. Except, of course, had that been my only course of action, I would have been up the creek without a paddle because XP wasn't capable of letting me talk to my network.
Mint, on the other hand, let me point out, had no trouble with any of my hardware. Thus Ubuntu-based Linux recognized the equipment, it set it up and let me get to work. It was Windows that proved to be a pain in the rump.
Greg Kroah-Hartman, a prominent Linux developer, is right. Linux Journal recently reported that he recently told an audience at the Ottawa Linux Symposium that "Linux supports more different types of devices than any other operating system ever has in the history of computing."
Linux isn't perfect that way, as Kroah-Hartman would be the first to admit. Based on what I experienced, though, Linux is much better than Windows at supporting modern hardware.
We have this illusion, that's just because Windows works on the systems it comes pre-installed on, that Windows has great built-in driver support. No, it doesn't. Once you move to installing Windows on a new system, you'll quickly find that Linux, not Windows, has the better built-in hardware support.
Yes, that's right. Linux, not Windows, is easier to install on a new PC. Just something to think about as you get ready to strip Vista off your new computer.
Oooooh, one year of difference. Except that it's not. Windows XP SP2 is not from 2001. It's from April 2008.
Hey youre the one claiming to be MS free for 10 years. Not my fault you dont actually mean it.
I don't buy Microsoft products. I have access to them through my job, just like any IT person in a sysadmin role would. But I don't buy them, I don't personally use them and I won't recommend them.
ANY device that is not recognized will pop up the New Hardware wizard. Thats what the wizard is for, if XP recognizes it
There you go with your assumptions again. Apparently, XP won't even recognize it.
And thats why we dont do hardware testing on virtual machines
I only used the virtual machine as an additional reference to the article (where real hardware was used) and to numerous web forums where users (using real hardware) registered complaints just like his.
Linux Mint - 2008
Windows XP Service Pack 3 - 2008
Windows service packs have never included drivers as standard issue. They might have new drivers and you can get them through the update site but they’re not in the SP unless the shipping drivers actually cause crashes, and that would only be a replacement not adding new. And there’s a good reason for that, people already whine the SPs are too big, how much bigger would they get if they added a bunch of possibly not needed drivers.
Windows XP drivers, regardless of SP - 2001
Not an assumption EXPERIENCE. If XP can see it, and it can’t automatically install drivers for it, and it hasn’t been told to ignore the fact that it doesn’t know what it is, then there will ALWAYS be a New Hardware wizard. 100% of the time ZERO exceptions. That’s what the New Hardware wizard is for ALL hardware that it can’t recognize.
Now if it can’t see it, like with your virtual machine, that’s a completely different situation. It can’t fail to recognize what it can’t see.
Gee, someone needs to tell Microsoft that. They seem to be under the impression that there are drivers on the SP CD.
On the other hand, they seem to like to delete drivers during the SP install, so that could be the cause of the article writer's problems.
Although, why Microsoft would just quietly delete the network driver for a widely used network card without warning the user is beyond me.
Now if it cant see it, like with your virtual machine, thats a completely different situation. It cant fail to recognize what it cant see.
No it's not. Windows has no idea whether it's running on real hardware or virtual hardware. The hardware is presented exactly the same.
It cant fail to recognize what it cant see.
Well, that part is true enough. Of course, it happens with real hardware too, just like it did to the writer of the article.
Bother to read the next sentence? Guess not.
Didn’t bother to actually read the article you linked to huh. It’s all about getting the OS to see the hardware. Like I said, if it can’t see it it can’t not recognize it.
Windows doesn’t know if it’s running virtual or real. But depending on your virtualization that’s going to keep a lot of hardware away from it. I work everyday with faxboards that Windows cannot see in any virtual environment, tried them all and none of them pass the existence of those boards through to a virtual Windows.
The faxboard issue is well documented on the VMWare and VirtualBox forums. Basically, it's not really a modem, it's a winmodem and the expects the host OS to provide most of the processing power. The guest OS tries to load it up and starts contention with the host OS.
Network adapters are not the same. To my knowledge there are no Windows network cards. They all work the same, thus they are presented to a real OS and a guest OS exactly the same.
If one OS can't see the hardware, but another OS can, then it's the OS that's the problem.
Nope it’s not a winmodem, and the faxboards I’m working with provide most of their own processing power, heck some of them even have cooling fans. Problem is the virtual software doesn’t actually pass all hardware through to the virtualized OS.
Never ever under any circumstances test hardware in virtual and think the test has even the slightest meaning. Simple rule, virtualization makes hardware access just different enough to completely invalidate the testing.
I've been busy installing a linux cluster for a university. Sorry for the late ping.
Cool. I've got a script running around here somewhere that allows me to multi-boot many different Live CD's off of one DVD. The disc boots up to GRUB, and you just choose which distro you want--boom--and it boots up.
I've installed up to 15 Live CD versions on one disc before.
There is a great deal of similarity in how the hardware is presented to the OS. The driver interfaces with the hardware at a very low level. This precludes accessing it from the guest OS since you can't have two operating systems accessing the hardware at the same time.
Never ever under any circumstances test hardware in virtual and think the test has even the slightest meaning.
Somewhat true. It depends on the hardware. Though I agree that testing using a virtual machine isn't sufficient.
For proper hardware testing, a real OS is required. Virtual machine access can identify other problems where the state of the real hardware is not properly passed to the guest OS, such as with the faxboards. Some USB devices have a similar problem.
However, Ethernet is a known quantity. The virtual Intel EtherPro passed to a guest OS is identical to a real Intel card as far as the guest OS is concerned. This makes it a good candidate for certain kinds of testing.
When I was considering replacing some older gigabit cards with new Intel cards I tested setting up VLANs and jumbo frames on the virtual cards first to determine proper syntax of the startup commands.
When I installed real cards with a real OS, the hardware probing results were identical.
Sweet!
Share?
:)
Even when the root OS isn’t doing anything with the hardware, even when the root OS doesn’t have the drivers installed so it CAN’T do anything with the hardware, it doesn’t get passed through to the virtual. The virtual never finds out they exist.
Nope always true. One of the first rules of testing: anything that sometimes invalidates the test ALWAYS invalidates the test. Is there hardware that works identically under virtual? Of course. But for the testing to be valid you have to prove that the hardware you’re working with does, which means you have to do at least one identical pass under virtual and real and get 100% identical results. Of course if you’re going to be doing real machine testing anyway there’s not much reason to do virtual.
There you go again assuming the hardware is identical to the function. Ethernet is a known quantity, Dell’s specific implementation of ethernet is NOT.
I’m Linux illiterate, and I’ve been getting by beautifully with Kubuntu for 3 years. Anything I have a question about, I head to a forum to look for the answer. I always come out having learned something new.
sono,
when is the last time you used linux and which distro? many of them now are as intuitive as XP
To broad a question...
After all if I said ‘is Unix for the masses?’ the first answer would be no way... but then you look at OS-X and its pretty clear some Unix is for the masses and some is not..
Linux is not all that dissimilar, I would not push some of the more tachie driven Linux distributions (like Damn Small Linux) on the masses but Ubuntu is as easy to use as windows..
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.