Slashdot Mirror
Longhorn: Fewer BSODs, More RSODs
Jan Theofel writes "Windows Loghorn will present you less BSOD. Joi Ito reports that Windows Longorn will get additional ROSD (red screen of death) for 'really bad errors.' So you will get less BSOD but some new RSOD. You can find a ROSD screenshot in a virtual machine in his weblog entry."
It's always nice to see Microsoft adding new "features". Now they can tout Longhorn's decreased BSOD occurrences. Although I'd think they'd wanna avoid red screens as they are angry colors.
Microsoft Sucks, F/OSS Rocks. I get mod points now right?
So where are the yellow and orange? Looks like MS has been taking advice from Tom Ridge.
I've not seen one of those in a long, long time.
Dont forget NetworkMirror :)
. ito.com/archives/2005/05/07/bsod_upgrades_to_rsod_ in_longhorn.html
http://www.networkmirror.com/adYJGbG8ajC3f55y/joi
I guess they've FINALLY fixed the last issue that casues a blue screen of death with Longhorn. :)
What exactly is a really bad error? I mean, a bad error versus a really bad error? That warrants a color change, anyways?
Frankly, I think customers ought to get rsod's for actually buying the damn product. That seems like a really bad error to me.
Now people will have heart attacks instead of just saying, "What the heck?" and getting frustrated.
Blues (and greens) are generally more soothing/comforting (which is why blue or green are most frequently favorite colors), whereas reds are more jarring (which is why it's used for stop signs, warning labels, etc.).
Well, here's a link to the pic... til I get slashdotted... http://209.193.18.52/RedScreen.jpg
Microsoft Sucks, F/OSS Rocks. I get mod points now right?
Longhorn is red-shifting... the release date must be receeding!
Since last night, I get this gem.
My own RSOD
As of 10/06/03, I hate COBOL developers.
If you are seeing BSODs almost daily, you either have faulty hardware or some seriously buggy drivers. Honestly folks, XP, and even 2000, BSOD very rarely.
Speaking of reliability, I was just thinking how Microsoft could reduce the complexity of the next version of Windows, Longtooth, due in 2009.
Longtooth will include a tremendous amount of new features implemented in completely new code. Many, but not all, existing features would be reimplemented in VisualBasic.NET just for the heck of it, even if mature versions are already implemented in C or C++. Programmers making the new VisualBasic.NET code would not be allowed to look at the code that already exists, so that new ideas might be better implemented. The features will be chosen by random for reimplementation.
All Microsoft code would assume that any Microsoft code (the OS and any Microsoft applications) is secure. This code will always execute with no checks to make it run faster. All other code will be subject to Longtooth's new security system, dubbed Microsoft Longtooth Security Center 2003. This feature will give users more control over processes that execute in their computers. I will explain some of its features here:
To maximize security, Microsoft Longtooth Security Center 2003 will make certain assumptions about the user. For example, users who use Microsoft products are assumed to know what they are doing. However, users of 3rd party applications not made by Microsoft are always assumed to be complete idiots. Therefore, all user interface events occurring outside of Microsoft applications will trigger a safety mechanism.
For example, each time the user moves the mouse in an area not controlled by a Microsoft application, the user will see crosshairs moving across the screen to indicate where the mouse will be located. When the user stops moving the mouse, an authentication window will appear and state: "The user has requested that the mouse be moved to the location on the screen indicated by the crosshairs. This area of the screen is controlled by untrusted code that may cause damage to your computer, your documents, or your network. Do you wish to allow the mouse to move to this location?" Buttons for "yes", "no", "details", and "help" will be displayed.
Selecting "no" will cause the mouse cursor to remain at its previous location. Selecting "yes" will bring up another window, requesting the user's password to authenticate the movement of the mouse. If the user enters the correct password, the mouse cursor movement will be authenticated to that user and the cursor will be placed at the new location. Selecting "details" will display the X and Y coordinates of the new position, followed by warnings against using untrusted rogue code such as Linux.
For additional protection, clicks, keys pressed on the keyboard, items selected in a menu, or other input events will trigger similar security mechanisms. Since Microsoft code is considered secure, these checks will not occur in windows owned by Microsoft code. Also, the mouse may be used to click on the above buttons and fields during mouse movement authentication. If any such movement of the mouse takes place during the authentication process, the mouse will still be moved to the location indicated by the crosshairs, but a bug in Windows will cause the cursor to immediately "bounce" back to the location where it was last used during authentication. Microsoft will refuse to fix the bug unless Linux is outlawed in all countries, even those countries that have no computers.
Many other authentication checks will be made by Windows. I'll return to this topic in a moment. First, let me mention that Clippy, the talking paperclip, along with other Microsoft characters, will appear during this proces
A mac then? :P
"Who says nothing is impossible? Some people do it every day!" - Alfred E. Neuman
Red screen reminds me of the infamous Amiga "Guru Meditation" error. I always said the Amiga was ahead of its time.
Comment removed based on user account deletion
is add a Green Screen of Death. Then they'll be able to add together death colors to get much needed functionality for TrueColor Screens of Death.
If you are seeing BSODs almost daily, you either have faulty hardware or some seriously buggy drivers. Honestly folks, XP, and even 2000, BSOD very rarely.
Exactly. I have never seen my XP machine at home BSOD, even when the video card was failing to the point that it was adding random horizontal lines across the display.
At work, I saw 2000 BSOD on several servers when we applied an MS hotfix that conflicted with some sort of secret kernel patch they'd given us a few years previously for those same machines.
I saw 2k bluescreen one other time, when a workstation had a zip drive and the user installed drivers for it from 1997 or so.
Other than that, the only time I've seen it happen is if I make an OS image on one machine and then try and use it on another with different hardware. That's still stupid, but at least I know how to avoid it.
This is in an environment with close to 1000 Windows servers and about 25,000 Windows workstations.
"...always new atoms but always doing the same dance, remembering what the dance was yesterday." -Richard Feynman
That's innovation for you!
Engineering is the art of compromise.
You brain dead moron. Microsoft takes advantage of the next generation directx 9 3d acceleration hardware available in all longhorn-supporting computers to blend the blue and red into a dark, rich purple using sophisticated algorithms, lighting effects, and large textures.
In case you were wondering, we know this because we stole a prerelease copy of Longhorn at musketpoint.
Is it related to this:
http://www.bbc.co.uk/wales/rsod/
Red screens are just too agressive. An error message already is very irritating, why the hell do we need it to be red ? I'd almost hate the Sarge installer because of that...
Ah.... I see that you are familiar with ATI's Catalyst drivers* then. Either that, or VIA's Hyperion drivers*. :-)
* Disclaimer: I use the term "drivers" very loosely.
Why, oh why, didn't I take the blue screen?
blog
How the hell do you expect to get a warning before your kernel crashes?
"WARNING: Your kernel will crash in ten seconds. Owing to the very nature of the event, there is nothing you can do about it."
there's more than one way to do me.
Edit your system.ini file as follows:
Under the [386Enh] header, add these two lines:
MessageTextColor=B
MessageBackColor=3
That will give you a bright cyan text on dark cyan background screen of death. Feel free to substitute other colors 0-F as desired. This works in 95, 98, and Me, at least. Red's in there somewhere - don't remember exactly where - just try a pair of values, wait the usual 15 minutes for a SOD, and see if you like the combination. I can honestly say I haven't seen a BSOD on my screen in months.
Who is John Cabal?
So there's actually a lot of BSODs that are 'preventative' in nature. That is, the kernel says "uh oh, that call should never have been made, the system *might* become unstable, shut it *all* down before any real damage is done".
Then there's "Boot disk not found", or "Boot disk failure", which are in fact real serious, because it's the end of the line for the machine.
Maybe they've broken down errors that are likely Kernel driver programming mistakes, and errors that indicate the system is severely damaged.
Actually, it isn't in the registry, it's in system.ini. I haven't been able to verify whether this works, as the computer I'm on hasn't had a BSOD since I got it. I take good care of it.
warning: This post is likely to contain gobs of dripping sarcasm. Consume at your own risk.
I've been seeing a lot less BSOD's in general since using 2K and XP.
In Windows 98, I would run the comptuer for a few days, and for no reason it would just start being slow and throwing random BSOD's at me.
Since using 2K and XP, I've seen a few recently, but they're all realated to a piece of faulty hardware that I've been too lazy to replace. Other than that, I can't reacall seeing a single BSOD in years on a computer of my own.
I'm honestly asking people. Have you run into BSOD's that really truely was 2000's/XP's fault instead of being some sort of hardware fuckup?
I'm the guy with the unpopular opinion