Windows: who designed this piece of crap?

[Spungo]

Situation 1:

You install Windows 7 or 8, and your computer's BIOS is in IDE mode. If you later realize: whoops I forgot to switch to AHCI mode, your Windows won't boot. You can switch back to IDE, boot into Windows and do some tweaking to get Windows to load the AHCI drivers, but it is a hassle.

What I find even more retarded is that you can't move hard drives between computers. Here's a true story. Let's say I have an old file server that is a bit slow. It has a legal copy of Windows on it, it has all of my programs configured the way I like. I don't want to screw with it. I just want to move to a newer computer. Guess what happens. It won't boot. It crashes. Absolutely retarded. MS should seriously consider firing most of its dev team. Or at least fire the ones that didn't quit due to stack ranking (a policy where employees are encouraged to sabotage the work of other employees to look better by comparison).

Linux:
No problems at all. Move the hard drive to another machine and it works fine. Time wasted: 0 minutes.

If MS wasn't retarded:
It should work, but it should deactivate the registration. Call them up, get a new key, away you go. Time wasted: maybe 10 minutes.

How it currently works:
Let's make this as slow and difficult as possible. Need to reinstall Windows. Download and install about 2gb of updates. Reset the computer multiple times while doing this. Need to reinstall all of your old programs. Need to reconfigure all of your old programs. Time wasted: about 1 day.

 

[mikeymikec]

How it currently works:
Let's make this as slow and difficult as possible. Need to reinstall Windows. Download and install about 2gb of updates. Reset the computer multiple times while doing this. Need to reinstall all of your old programs. Need to reconfigure all of your old programs. Time wasted: about 1 day.

As things currently are, I suspect one could get away with installing Windows 7 on say AMD kit then moving the drive to an Intel platform as long as one doesn't install any third party drivers. If one does, then the worst case scenario is that the destination machine won't boot, the best case is that those drivers then need removing and the proper ones installed.

Is this really a Windows problem per se? What if AMD or Intel write something into their chipset drivers based on the (reasonable, up to a point) assumption that their software will only be used on an AMD platform, but if the job was to be done as properly as possible, there would be a verification of matching hardware during driver initialisation rather than a potential assumption that the hardware on day 1 is the same as on day 720.

You're saying that Linux doesn't have the problem, but a lot of the time the hardware compatibility work isn't done by the same kind of people AFAIK as on Windows. For starters there's code review a lot of the time (even if Intel or AMD contributes code) and closed source is frowned upon for Linux and similar platforms, which encourages better coding practices.

---

Regarding the original topic, I would like to have seen the BSOD outputs in the scenarios that the OP described. I bet it's a third party driver that's thrown the spanner in the works.

 

[imagoon]

How it currently works:
Let's make this as slow and difficult as possible. Need to reinstall Windows. Download and install about 2gb of updates. Reset the computer multiple times while doing this. Need to reinstall all of your old programs. Need to reconfigure all of your old programs. Time wasted: about 1 day.

Don't tell that to my current desktop. It started as a laptop on spinning disk, disk imaged to an SSD, laptop -> q6600 desktop -> i5 desktop. Maybe it forgot to screw up?

Time spent each time was less than an hour. Longest one was when the NIC on the i5 wasn't natively supported so I downloaded it to my phone and then USB'd it over.

 

[imagoon]

As things currently are, I suspect one could get away with installing Windows 7 on say AMD kit then moving the drive to an Intel platform as long as one doesn't install any third party drivers. If one does, then the worst case scenario is that the destination machine won't boot, the best case is that those drivers then need removing and the proper ones installed.

Is this really a Windows problem per se? What if AMD or Intel write something into their chipset drivers based on the (reasonable, up to a point) assumption that their software will only be used on an AMD platform, but if the job was to be done as properly as possible, there would be a verification of matching hardware during driver initialisation rather than a potential assumption that the hardware on day 1 is the same as on day 720.

You're saying that Linux doesn't have the problem, but a lot of the time the hardware compatibility work isn't done by the same kind of people AFAIK as on Windows. For starters there's code review a lot of the time (even if Intel or AMD contributes code) and closed source is frowned upon for Linux and similar platforms, which encourages better coding practices.

---

Regarding the original topic, I would like to have seen the BSOD outputs in the scenarios that the OP described. I bet it's a third party driver that's thrown the spanner in the works.

Moving from AMD to Intel or back is fairly easy. The drivers for the drive controller generally is the only thing that is needed to be installed ahead of time. The other easy way is if both support "generic IDE" you can install the IDE drivers, flip the BIOS to IDE mode move over, install the correct drivers and flip back. This does require you to flip the enabled bit in the registry for the driver controller drivers to make it smooth but that is documented on countless websites including MS's own site.

 

[Spungo]

You're saying that Linux doesn't have the problem, but a lot of the time the hardware compatibility work isn't done by the same kind of people AFAIK as on Windows. For starters there's code review a lot of the time (even if Intel or AMD contributes code) and closed source is frowned upon for Linux and similar platforms, which encourages better coding practices.

Linux:
You appear to be using an AMD motherboard. I'll use the AMD drivers to start the computer.

Windows:
Let's make no attempt to detect anything. Bios set to ACHI? Won't boot. Change motherboard? Won't boot. A butterfly in China flaps its wings? Won't boot.

 

[mikeymikec]

Linux:
You appear to be using an AMD motherboard. I'll use the AMD drivers to start the computer.

Windows:
Let's make no attempt to detect anything. Bios set to ACHI? Won't boot. Change motherboard? Won't boot. A butterfly in China flaps its wings? Won't boot.

Windows actually does detect quite a lot, and in some respects Windows has run rings around almost every other OS in this respect, it's not just very tolerant when it comes to detecting storage to boot from. IMO a lot of crapness in Windows revolves around how it handles storage, it's quite inflexible in this respect.

Every platform has advantages and disadvantages when compared to its competitors.

 

[imagoon]

Linux:
You appear to be using an AMD motherboard. I'll use the AMD drivers to start the computer.

Windows:
Let's make no attempt to detect anything. Bios set to ACHI? Won't boot. Change motherboard? Won't boot. A butterfly in China flaps its wings? Won't boot.

Enable the disk driver scan at boot up and you would be fine. This is a well known issue even discussed in this very thread. You can switch system boards with zero issue as long as you take a tenth of a second and enable the driver rescans.

 

[jcgriff2]

You boot off the CD and start the install, then half way through the install process it wants to load drivers for the cdrom drive! WTF? I BOOTED OFF THE CDROM DRIVE! Silly Windows.

My guess would be a corrupted DVD disc or a bad DVD drive -- or the installation itself was not going well (corrupted).

cdrom.sys is on every Windows installation media and Windows would never need to ask a user to provide it.

 

[vailr]

AFAIK, all Intel 7 series chipsets (such as Z77) & later default to AHCI mode in bios.
I do agree that Microsoft should focus some attention on allowing more ease of switching between IDE mode & AHCI mode. They could at least: make it equal to Linux in that respect. They're currently accepting suggestions for improvements in Windows 10, so...maybe there's an opportunity?

 

[ZippyDan]

I am a little confused by this comment. You have been updating BIOSes for 25 years, and you are just now noticing that you can't change drive communication modes without Windows caring?

Many of our computers were still running old installations of Windows XP and Windows Vista until recently, and as such were running BIOS that either did not support AHCI or were still in IDE mode. I just did a huge maintenance and update program on many computers in many offices around the globe. All computers in the company are now running Windows 7 or Windows 8. In the process of updating, I tried to switch as many of the motherboards to AHCI mode as possible. I was also surprised that Windows 8 still does not handle this change smoothly. It is my first time working with Windows 8 so extensively.

 

[us3rnotfound]

I am a Windows Defender ., just kidding but yeah I see your point OP. I have been using Win since 95 days, and very recently had a hell of a deal with my Win8.1, I can't even explain what the hell happened, it totally just broke itself from the inside like a cancer or something, really bizarre, couldn't fix it so I needed to reformat. It was obviously some bizarre .NET issue, couldn't fix via DISM restore-health or SFC/scannow commands.

With Windows, as I've progressed in my technical abilities, I've found before any major installs I sit and plan out everything before I go in because of stuff like you say with the AHCI mode. I cannot count how many times I've ended up in a place I didn't want to be so I had to wipe the slate clean and get it right for the last time. Lol but I'll still be using Windows next year because, well what choice do I have.

 

[Red Squirrel]

My guess would be a corrupted DVD disc or a bad DVD drive -- or the installation itself was not going well (corrupted).

cdrom.sys is on every Windows installation media and Windows would never need to ask a user to provide it.

Nope I've seen this many times on many different machines using many different CDs. If it does not like the cdrom drive at that stage of the install it will prompt to load a driver. It seems to happen mostly in vm situations. It seems fine in ESXi but I think it was either Qemu or KVM that I saw it happen.

 

[imagoon]

Nope I've seen this many times on many different machines using many different CDs. If it does not like the cdrom drive at that stage of the install it will prompt to load a driver. It seems to happen mostly in vm situations. It seems fine in ESXi but I think it was either Qemu or KVM that I saw it happen.

There was an era where the some of the VM developers thought they knew better than the rest of the world and decided to not be fully ATAPI compliant. During that time you did need a driver because they had written an a virtual device that needed drivers. Hardly a "Windows" issue.

 

[mikeymikec]

My guess would be a corrupted DVD disc or a bad DVD drive -- or the installation itself was not going well (corrupted).

cdrom.sys is on every Windows installation media and Windows would never need to ask a user to provide it.

Or faulty RAM, I've seen file transfer errors during setup a few times because of it.

 

[Red Squirrel]

There was an era where the some of the VM developers thought they knew better than the rest of the world and decided to not be fully ATAPI compliant. During that time you did need a driver because they had written an a virtual device that needed drivers. Hardly a "Windows" issue.

But windows still saw the drive to be able to boot off it and read the data, so why does it suddenly decide half way though that it can't read it anymore? It could just do whatever it was doing before to read it. A cdrom drive is not something that should require drivers.

 

[ninaholic37]

I am a Windows Defender .

When I read that I thought you were saying that you were the protection program that comes with Windows and your PC Tower avatar was talking. User Not Found means that you are the computer :awe:

 

[imagoon]

But windows still saw the drive to be able to boot off it and read the data, so why does it suddenly decide half way though that it can't read it anymore? It could just do whatever it was doing before to read it. A cdrom drive is not something that should require drivers.

That is a different technology called floppy emulation that stops working when the BIOS no longer has control of the device like when Windows initializes the hardware.

CDROM drives certainly require drivers. If you are using something that doesn't follow the ATAPI standards, it will require drivers because cdrom.sys will be unable to operate it. Even today if you plugged in one of those old Sound Blaster Pros with a 2 X CDROM on it, you would need drivers. The reason is that the Sound Blaster Pro CDROM was not ATAPI. Same issue with Firewire CDROMs. Granted Windows includes the fireware cdrom driver assuming the firewire driver itself is found.

Just because you state that they shouldn't need drivers, doesn't make it so or any less reality.