OT: Yet Another Unhappy Customer for Vista

JackShephard wrote in news: snipped-for-privacy@4ax.com:

Nice cap-locked foam. It's a wish. Why would you "wish" harm on me?

just

You did. If it really mattered, I'd remember what it was, exactly. But since it came from you.. pffffffffft..

Crash Street Kidd wrote in news: snipped-for-privacy@i38g2000prf.googlegroups.com:

FOADIAF

innews: snipped-for-privacy@i38g2000prf.googlegroups.com:

Shut up, Osterwald!

CSK

Darn it. Stop morphing you worthless anklebiting git of a troll and STAY in my filters.

Sit boy. Stay.

Please do not distribute your cancellable spam to AUK. TIA.

Were you always easy to troll, or is it a recently acquired talent?

Peter J Ross wrote in news: snipped-for-privacy@pjr.gotdns.org:

I have reported you. You auk kooks are retards.

Go f*ck yourself, MassivePong.

-- \___ Proud Cog #1 in the AUK Hate Machine _(AUK)====:: Do *you* think that you have the Right Stuff? /='='='='-, Apply TODAY by addressing a gratuitously cruel (O+O+O+O+O) flame to: "Uncle Fester", C/O soc.singles & AUK. ~^^^^^^^^^~~~^~^^~'~~^'^~~~"~~'"~^~'"~~^~"~'~^'^~^~^^~^~"~^~"'~'"~^~~

With the emphasis on the word "should" ;-)

For performance reasons: Not bothering to actually commit data to a hard drive (or other non-volatile storage) until absolutely necessary is a big performance win in a file system, since the (very limited, relative to system memory) bandwidth of the hard drive can then be used for servicing cache misses. Also, telling the OS to shut down gives the system a chance to tell all of its peripherals to go to sleep or power themselves down -- a hard drive or video camera or whatever on a USB bus otherwise doesn't "know" it's supposed to power down just because you killed the power to your PC.

If you think about it, a PC with a cheap UPS (that tells the system to shut down) isn't all that different from any embedded piece of hardware that has a power fail interrupt and a big enough filter cap (or whatever) to keep it going for some milliseconds (or seconds) after power fails -- it's just that the PC is operating on the time scale of many seconds (or minutes :-) ).

I imagine the official MS response would be that you need a UPS. :-)

That's kinda comparing apples and oranges -- I think the folks at Microsoft and QNX would both agree that, in *most* situations, they don't end up as competitors.

Although I suppose that's a lot more naive individuals trying to use Windows as an RTOS than QNX as a general-purpose desktop OS.

Yep, it's easier to require the customer to devise and pay for a somewhat fail-safe strategy than spending time to ruggedize the software.

Which QNX does nicely BTW except that there are hardly any apps written for it. They had a little web app coming with it as a sample and the only browser that ever came close to its solid behavior was Mosaic.

Understood. But that is still no reason why an ordinary app should crash in a way that you can't restart it after restoring power. Everyone would accept data losses, but not a hard freeze.

Then their designers need to get back to the drawing boards.

Well, in medical we would (in most cases) end up with some serious egg in the face if we designed UPS-reliant. Which is why our team had to spend a whole lot of time re-writing low level routines that were, ahem, "sub-optimal". Basically doing the job some OS designers should have done. Guess what, you could just pull the power plug on that machine and it would come right back up the next time.

I agree with that. I suspect that many applications can be killed (caused to freeze up) if given enough corruption in their data files, and at the very least there should be a documented means of resetting the configuration data files if need be to get things going again. I've had similar experiences to yours (albeit not due to power failures, AFAIK) where Office gets confused and it seems quite impossible to get it working again, since uninstalling Office doesn't actually uninstall *everything*.

I've also managed to munge up OpenOffice once, but there I *was* able to find documented (on the web) how to delete the corrupt configuration files. Score one for OpenOffice...

In the case of USB, I'm not certain what a device is "supposed" to do, but I can see the USB desginers thinking, "hmm... well... a hub *could* detect a loss of connection to a PC, and it *could* power down everything connected to its ports, but if one of those devices happens to be a hard drive, that really might not be the best idea..."

Sure, and that's perfectly reasonable for medical equipment, but I think you can make a decent argument that it's better for, e.g., Amazon.Com or Google.Com or anywhere else that has an insastiable appetite for huge and fast databases that it can be better and cheaper to add a UPS than accept the performance hit that not requiring a UPS would impose and adding extra equipment to compensate.

It really is a problem that's driven by people wanting PCs to be fast and cheap first and reliable second in most cases... whereas for medical devices the priorities are different. "Good, fast, and cheap -- pick two" seems to apply here...

---Joel

Say Joerg,

What do you think of Google's approach to purposely getting cheap, not-so-reliable hardware?

-->

I know how much you like making things cheap AND reliable, after all. :-)

---Joel

I am no longer using OO except for transferring stuff to pdf. I found it to be really slow, bloated, possibly fraught with memory leaks and it delivered the record of tough crashes to date. All on various machines.

Have to admit, MS-Word is more reliable here.

Back around 1600 or so they would have locked the designers of such HD into a basket, hung it over the walls of the local fortress and dunked him into the moat a few times.

Well, sure, those are huge volume sites. But for a small office/home? In the early 90's nobody had a UPS. Because the designers back then must have been smarter and nobody needed it. On my DOS machine the manual said I could just flick the big red power switch. Which I did every day without any negative conesequences.

Ideally they shouldn't be different. Why is it that a bunch of (good) SW engineers in California can do it, while others that are supposed to do it can't? It's not that this machine was any slower in the end, it was just better.

Redundancy is one thing but someone has to sit down as calculate the cost for all those maintenance workers. Not so sure if the number at the bottomline would be in the green.

I believe John Larkin's approach is better, find a good quality PC with RAID drives, negotiate a nice deal, and then buy tons of the same. IOW a fleet purchase.

ALL modern drives fully park.

The "Windows" that is used as an RTOS is not anything like the consumer or business PC/server level products. It likely more closely resembles the embedded offerings.