View Single Post
  #7  
Old April 10th 19, 03:53 AM posted to microsoft.public.windowsxp.general
Paul[_32_]
external usenet poster
 
Posts: 11,873
Default Anti blue screen protecter

Bill in Co wrote:
J. P. Gilliver (John) wrote:
In message , Paul
writes:
JJ wrote:
On Sat, 6 Apr 2019 12:39:13 -0700 (PDT), AK wrote:
Does anyone use a anti blue screen protector?

They are rather expensive.

Thanks,
Andy
Not worth it, IMO. Because it doesn't solve the problem which causes
the
BSOD in the first place. It's like ignoring a court order that'll
always end up in a bad way.
There's two ways to do what I think he is referring to.

[learned discourse on backlights deleted!]

My first thought was that this was an April Fool thread, but the initial
post appears to be dated the 6th.

But I read it as referring to some piece of software that prevents BSODs
- *or claims to*. I've never heard of such, but I can just about believe
it as possible. It would have to be a very complex piece of software,
though - it would more or less have to sandbox or VM your entire system.
At the very least, I imagine it would have a significant effect on
performance.


That's the way I read it too! If such a thing existed, it would have to
operate at a very low level, and run all the time in the background,
monitoring everything. But I doubt it does exist - at least for
intercepting blue screens. I seem to vaguely recall some utility program
that could run in the background and intercept some stuff that went astray
at the higher levels, however, but can't recall the program name now


If such a thing existed... um, there'd be a patent.

If you hark back to your MacOS days, which was a cooperative
multitasking system, that would illustrate what a difference
a good OS partition makes. That OS used to crash like crazy
(once or twice a day, I used to keep MacsBug loaded to try to "escape").

g Finder

And it only got worse going from MacOS 6 to MacOS 9. *Every part*
of the environment, *every* application program (MSWord!),
plays a part. As the thread of execution winds its way through
all that code sequentially, many times a second. Any little
leak, a burp or a fart, could spell catastrophe. Each program
"gives up control" ("cooperates") after some period of milliseconds
and says "here, you use the processor for a bit while I sleep".
Programs have to be polite (and give up regularly), to keep
such abominations responsive.

All the major OSes eventually got preemptive multitasking.
The kernel lives in its own "ring", for protection from the
elements. It "remotely controls" (pre-empts) the execution
of processes in userland. If Firefox dies, nobody cares,
least of all the kernel. If Firefox dies, you don't reboot
the computer, you just restart Firefox, then curse Firefox
for being "bad software" (no matter what the real reason
might have been). The kernel in such cases, can have
extremely long up-times (at least, if we're using a
cooperative tasking OS as our baseline).

The quality of the preemptive implementation, how
many "land mines" you put in it, determines how close
to ideal it is. Microsoft likes to poke fun at NVidia
and ATI drivers, as a major source of instability. You
can be sure that on Solaris (where the hardware manufacturer
has something to say about video hardware), the opportunity
for bulletproof hardware is a lot higher.

And these exposures can even be seen in an "overclocking"
situation. If you run Linux on a processor that is clocked
higher than it should be, the first thing that disappears
is programs. You see Firefox crash. You see the file manager
crash. You see decorations or dialogs crash (ones that
have their own process to run them). The very last thing
you see is the kernel panic, when it comes crashing down.
Even that kind of a demonstration, helps show how a "small"
kernel is less suspectable to mischief than a bloated "Firefox".
Even if this is not really all that logical of a demonstration,
it neatly shows how a different partitioning can make the
"core part" of the OS stable enough, you don't need to keep
MacsBug loaded :-)

In light of the apparent good underpinnings of modern OSes,
it would be "snake oil" to be selling a "BSOD preventer". The
design itself is a "BSOD preventer". Rewriting the NVidia
driver might be as close as you get to "BSOD prevention" :-)
(At least if you inhale some of those MS statements on the topic.)

Paul
Ads