So much for no kernel change - "Windows 7 *is* Minwin


Recommended Posts

Oh god. It's back. The last time this thread popped up (A couple of weeks ago), it was mostly post after post of people showing off just how much they don't know what they're talking about...

Edited by MioTheGreat
Link to comment
Share on other sites

Oh god. It's back. The last time this thread popped up (A couple of weeks ago), it was mostly post after post of people showing off just how much they don't know what they're talking about...

And of course you do, right?

Link to comment
Share on other sites

NT Kernel still suffers from BSOD mostly caused from overclocking and some bad drivers. I think MS can do something about bad drivers. Don't let non WHQL to be installed. In case of hardware failure, try to shut it off and notify user unless it's mobo or memory related.

Link to comment
Share on other sites

And exactly how is the microsoft SOFTWARE supposed to fix hardware level error crashing your computer because you overclocked it ? magic ?

as for bad drivers, they allready did most of what hey coudl here by moving graphcis driving to user level. and requiring certification. frankly without causing severe performance loss you'll never be able to make a system where a hardware driver can't EVER crash the system. but if you stick to old reliable drivers instead of allways jumping on the latest ones, you're pretty safe anyway.

Link to comment
Share on other sites

NT Kernel still suffers from BSOD mostly caused from overclocking and some bad drivers. I think MS can do something about bad drivers. Don't let non WHQL to be installed. In case of hardware failure, try to shut it off and notify user unless it's mobo or memory related.

Don't allow non-WHQL drivers? That's unlikely. There are enough complaints about the code signing requirement for kernel-mode drivers on x64. If they were required to go through WHQL it'd be even worse. Besides, WHQL doesn't prevent a driver from crashing. It helps, but most of the time the big companies like Nvidia find ways around the requirements by having a "normal" and "performance" mode, where only the normal mode is tested by WHQL.

If bad hardware results in corrupted memory or calls something from the wrong IRQ level it's too late. You can't just turn that piece of hardware off, because the damage is already done.

Link to comment
Share on other sites

Don't allow non-WHQL drivers? That's unlikely. There are enough complaints about the code signing requirement for kernel-mode drivers on x64. If they were required to go through WHQL it'd be even worse. Besides, WHQL doesn't prevent a driver from crashing. It helps, but most of the time the big companies like Nvidia find ways around the requirements by having a "normal" and "performance" mode, where only the normal mode is tested by WHQL.

If bad hardware results in corrupted memory or calls something from the wrong IRQ level it's too late. You can't just turn that piece of hardware off, because the damage is already done.

If company can't get WHQL drives through the doors, well it's probably utter garbage. This way at least MS can reduce the problems. As far as IRQ i agree, still i think there are certain situations where it could OS find its way out. I think original IBM PC concept is the worst IT industry could come up with, but it's too late now...that's for another discussion.

And exactly how is the microsoft SOFTWARE supposed to fix hardware level error crashing your computer because you overclocked it ? magic ?

as for bad drivers, they allready did most of what hey coudl here by moving graphcis driving to user level. and requiring certification. frankly without causing severe performance loss you'll never be able to make a system where a hardware driver can't EVER crash the system. but if you stick to old reliable drivers instead of allways jumping on the latest ones, you're pretty safe anyway.

They can't fix overclocked systems. I didn't say they could. I just said what was one of the reasons, but again people who overclock are advanced users and so BSOD is not problem for them...it's something to expect. In other words i said that NT Kernel matured and it's very stable.

Link to comment
Share on other sites

NT Kernel still suffers from BSOD mostly caused from overclocking and some bad drivers. I think MS can do something about bad drivers. Don't let non WHQL to be installed. In case of hardware failure, try to shut it off and notify user unless it's mobo or memory related.

You're kidding, right? Say goodbye to beta Forceware drivers and tweaked sets if everything MUST be WHQL'ed. There's also plenty of drivers on my system that do not induce any BSODs whatsoever, yet they aren't WHQL'ed. (i.e. drivers written by freeware authors)

Link to comment
Share on other sites

You're kidding, right? Say goodbye to beta Forceware drivers and tweaked sets if everything MUST be WHQL'ed. There's also plenty of drivers on my system that do not induce any BSODs whatsoever, yet they aren't WHQL'ed. (i.e. drivers written by freeware authors)

True, but you do gotta remember that most of (like at least half) of the problems reported about windows involve some sort of 'tweaked' driver or setting...which isn't MS's fault is it?

Link to comment
Share on other sites

True, but you do gotta remember that most of (like at least half) of the problems reported about windows involve some sort of 'tweaked' driver or setting...which isn't MS's fault is it?

Not necessarily. Where did you pick that up from?

Tweaked sets are nothing more than simple ini/inf/cfg changes, no binary code is patched at all. Sometimes the authors of tweaked sets do mix and match binaries from older driver sets.

What this guy is suggesting (well, most of said suggestions are terribly outlandish but meh) is even worse than MS blocking out non-signed drivers in 64-bit Windows.

Link to comment
Share on other sites

Not necessarily. Where did you pick that up from?

Tweaked sets are nothing more than simple ini/inf/cfg changes, no binary code is patched at all. Sometimes the authors of tweaked sets do mix and match binaries from older driver sets.

What this guy is suggesting (well, most of said suggestions are terribly outlandish but meh) is even worse than MS blocking out non-signed drivers in 64-bit Windows.

Sorry, I actually left out 3rd parties (I.e nVidia, ATI, anyone else who makes drivers)

Lol, so MS is suggestign that anyone without WHQL is denied driver access? does it cost to get this?

Link to comment
Share on other sites

It's pretty sweet eh. What really annoys me is people flaming MS because 7 is built on Vista, but in reality, XP was built on Windows 2000 as well. It's the same thing, yet people are going on about how it should never be done

Also, lest anyone forget, Windows NT 3.1 had similar requirements to Windows 3.x enhanced mode, except that you couldn't run NT on a 386SX, and it required 8 MB of RAM, rather than four. (At one point I dual-booted Windows 95 SR2 and NT 3.51 on a 386DX-40 (AMD) with 16 MB of RAM.

It was NT 4 that dropped support for the 386 (only because the Pentium had become mainstream).

As far as the major changes in kernels over NT's lifetime, it's actually comparable to the changes in the Linux kernel from 2.2 to today (2.6.27.x). 2.2 is comparable to NT4, 2.4 to NT 5, and 2.6 (today) would be the equivalent to NT 6. However, look at the changes *just in 2.6* and you get an idea of the changemap between Vista and 7.

Link to comment
Share on other sites

It's pretty sweet eh. What really annoys me is people flaming MS because 7 is built on Vista, but in reality, XP was built on Windows 2000 as well. It's the same thing, yet people are going on about how it should never be done

The people who do such flaming generally don't understand the first thing about operating system. These are the kind of people who judged Vista solely by its GUI, not understanding all the major changes that occurred under the hood.

Vista was a much better kernel model than XP, but it also ushered in a new method of building operating systems for Microsoft: build the kernel, then cut the cloth and stitch it over each client and server release. Whereas XP was built on top of Windows Server 2000, Windows Vista and Windows Server 2008 are virtually identical, except the latter has certain server functionality. This is more similar to how Apple builds its client and server releases. And Windows 7 will continue this trend, also being almost an "off-the-shelf" version of Windows Server 2008, except with end-user features being added, instead.

Link to comment
Share on other sites

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.