[HN Gopher] Why is there a screen that says "It is now safe to t...
       ___________________________________________________________________
        
       Why is there a screen that says "It is now safe to turn off your
       computer"? (2016)
        
       Author : susam
       Score  : 35 points
       Date   : 2025-02-16 20:08 UTC (2 hours ago)
        
 (HTM) web link (devblogs.microsoft.com)
 (TXT) w3m dump (devblogs.microsoft.com)
        
       | gnabgib wrote:
       | (2016) Discussion at the time (99 points, 44 comments)
       | https://news.ycombinator.com/item?id=11527671
        
       | accrual wrote:
       | These shutdown splashes are one reason why I enjoy working with
       | AT motherboards, those with the old P8 + P9 power connector ("red
       | to red and it's dead").
       | 
       | There's some charm and nostalgia in the computer only having two
       | states - on and off. No standby power, no S1-S4 states, no wake-
       | on-LAN, no management engine, etc. Press the power button -
       | hopefully a chunky rocker switch - and the system is _off_ (minus
       | the RTC battery).
       | 
       | I have an XP machine running on such a motherboard and it too
       | supports the "safe to shutdown" splash. Hibernate also works, so
       | if I really want to, I can put it to "sleep" with no power at
       | all.
        
         | wolrah wrote:
         | > Press the power button - hopefully a chunky rocker switch -
         | and the system is off (minus the RTC battery).
         | 
         | My favorite part of this era were those under-monitor power
         | switch gadgets, so you could just leave the AT power switch on
         | as well as the printer, monitor, modem, etc. and have one
         | master power switch that turned everything on or off.
         | 
         | I still have an early '90s IBM PS/1 that was my family's first
         | computer. The only accessories that survived from then are the
         | CH Flightstick and Gravis Gamepad, the original keyboard,
         | mouse, and monitor are long gone (I really regret not having
         | the wonderful original clacky IBM keyboard), but it still works
         | and still boots off its original 129MB hard drive.
         | 
         | If I find one of those power switch things at Goodwill or
         | whatever I'll pick it up without a second thought.
        
         | dist-epoch wrote:
         | > hopefully a chunky rocker switch
         | 
         | Because wall-power was routed through that switch, quite
         | dangerously, the cable could get pinched in the case.
        
         | jwineinger wrote:
         | May I ask why you still have an XP machine running?
        
         | 1970-01-01 wrote:
         | I wonder if one forcefully disables all ACPI support and
         | drivers, are they allowed to go back to the binary ON/OFF
         | lifestyle on a modern x86 system? What does Windows do?
        
       | m348e912 wrote:
       | The issue I have is the wording.
       | 
       | The problem really comes down to the fact that windows couldn't
       | actually power off the computer like it can today.
       | 
       | I think the message should have said: "Windows shutdown complete.
       | Please power off the computer."
       | 
       | This more clearly implies that not only is it safe to turn off
       | your computer, it's recommended you do so.
        
         | dist-epoch wrote:
         | > Please power off the computer.
         | 
         | Sounds a bit like ordering the user, I guess that's why they
         | went with the other message.
        
         | advisedwang wrote:
         | This message helps reinforce that shutting down the computer
         | before seeing that message was NOT safe.
        
         | bombcar wrote:
         | Remember that it's an all-black screen that came on after a
         | blue cloud screen with "Please wait while your computer shuts
         | down".
        
         | davidt84 wrote:
         | The user needed no encouragement to switch it off.
         | 
         | They'd spent the last 10 minutes swearing at windows for taking
         | forever to shutdown and were about 10 seconds from saying screw
         | it, and turning it off anyway.
        
       | xg15 wrote:
       | Semi-OT, but is there still any situation in modern computers
       | where turning it off would be "unsafe"? I.e. if you yanked the
       | power cord at some specific time, the system would take
       | irreversible damage?
       | 
       | Was this actually a real danger at all at some point?
       | 
       | It's obvious that cutting the power without shutdown will lead to
       | data loss of everything not written to persistent storage. But
       | for old PCs, the warning was always that it could also physically
       | damage the PC, e.g. damage the hard drives or something like
       | that. Is this still the case?
        
         | BugsJustFindMe wrote:
         | > _But for old PCs, the warning was always that it could also
         | physically damage the PC_
         | 
         | I don't remember this being true. I remember it always being
         | about data loss.
        
           | dist-epoch wrote:
           | Older hard-drives didn't automatically park the head on power
           | loss, leaving it in a dangerous position above the platters.
           | Any shock would smack it into the platters.
        
             | SoftTalker wrote:
             | My first hard drive came with a utility program to park the
             | heads. Was recommended to run it before powering off
        
         | Ekaros wrote:
         | I think it could be harm to components when you are writing in
         | their non-volatile memory. Namely at least motherboard and GPU.
         | Both often have flashable memory for BIOS/UEFI and VBIOS(for
         | GPU). Interrupting power at wrong time when these are being
         | written could result in non-functional components. Some models
         | do offer dual-bios which lowers the risk.
        
         | rzzzt wrote:
         | SSDs without a capacitor bank that keeps the lights on for a
         | bit more time after poweroff can leave their internal data
         | structures (flash translation layer) in an inconsistent state.
         | The effect ranges from nothing to data loss to bricking.
        
         | advisedwang wrote:
         | It was less about physical damage, and more about leaving the
         | system in a state that couldn't boot up. In particular,
         | incomplete writes leading to corruption of the windows registry
         | was a frequent problem.
        
         | Syonyk wrote:
         | Yes, in some cases. The common case for this would be writing
         | the flash chip ("Updating the BIOS"). _Some_ boards are more
         | cautious about this than others, and a lot of  "enthusiast"
         | (overclocker) boards have dual BIOS chips, so you can always
         | swap to a working one and then reflash the corrupted one. But
         | very few modern systems will tolerate a power cut in the middle
         | of a firmware update very well, and even if it does, it's often
         | in a "Well, now, create this particular USB stick with exactly
         | this filename on a FAT16 partition, boot it with it plugged
         | into this particular USB port, and it might recover things"
         | sort of method.
         | 
         | You'll notice most laptops refuse to update their firmware
         | without the battery reasonably fully charged, _and_ the power
         | adapter plugged in. This is why.
         | 
         | And then we get into the definition of "irreversible." Yes, you
         | can recover from this with an external SPI flasher, if you can
         | get the firmware image into the proper state for the SPI flash,
         | or have an image from before. But most people don't have this
         | option.
         | 
         | As for hard drives, I'm of the impression that most newer
         | drives will handle a power cut by using the inertia left in the
         | platters and the motor-as-generator long enough to drive the
         | heads back to a safe position, but I've not done any sort of in
         | depth experimentation on it. A drive with a parking ramp for
         | the heads (or a parking track) has those things for a reason,
         | and "scraping the coating as the drive spins up somewhere in
         | the data section" isn't good for drive longevity.
        
           | jmholla wrote:
           | > But very few modern systems will tolerate a power cut in
           | the middle of a firmware update very well, and even if it
           | does, it's often in a "Well, now, create this particular USB
           | stick with exactly this filename on a FAT16 partition, boot
           | it with it plugged into this particular USB port, and it
           | might recover things" sort of method.
           | 
           | I recently did this with one of my Thinkpads, doing it during
           | a battery firmware update among others. I wasn't successful
           | in finding such a solution though. Where have you had success
           | in finding some of the more obscure repair techniques?
        
         | jowea wrote:
         | I think you can have problems today if you have a power failure
         | mid motherboard firmware/BIOS/UEFI upgrade. Some mobos I heard
         | have some protection or some way to unbrick it but it could be
         | dangerous.
         | 
         | And it was a real danger before for the running sofware system,
         | I think modern filesystems are more resilient against this sort
         | of problem.
         | 
         | Never heard of actual physical damage though, but I didn't live
         | that era much, can just barely remember the screen in TFA.
        
       | ryandrake wrote:
       | Yes, I know about disk caches and data corruption and blah blah
       | blah. But to me, it's yet another condescending reminder of who's
       | actually in charge of the computer (hint: It's not the user, like
       | it was in the days of yore). You don't just turn off your
       | computer anymore. No, that would put the user too much in
       | control. Now you _request_ to the operating system:  "Pretty
       | please, with sugar on top, if you're not too terribly busy doing
       | whatever it is you're doing, would you mind powering off?"
       | 
       | The computer has its own agenda and agency, and increasingly the
       | user's wishes are unimportant. We've been conditioned to _ask_
       | the computer to do things rather than _command_ it.
        
         | dullcrisp wrote:
         | This is why I alias kill to kill -9 on all my systems.
        
         | gapan wrote:
         | I don't think it has anything to do with the "good old times".
         | It is a problem that's introduced as soon as a storage medium
         | is used. Sure, you could shutdown your C64 anytime, but even
         | then, if you did it while it was writing to the tape/floppy,
         | then your data was lost. And I definitely remember having to
         | park the hard drive on an XT before shutting it down.
        
         | kragen wrote:
         | It's a good reminder that the days of yore were mostly actually
         | pretty terrible for user control, because we're talking about a
         | problem that currently popular computers like this cellphone
         | and any Linux box with ext4fs no longer have. And of course
         | before the move to Microsoft Windows around 01991 an awful lot
         | of people were using minicomputers and mainframes they not only
         | weren't allowed to turn off, but often weren't even allowed to
         | see at all. A lot like AWS.
        
       | 1970-01-01 wrote:
       | >This message was displayed when shutdown was complete and the
       | computer did not support software-initiated power-off. It told
       | the user, "Okay, shutdown is complete. It's okay. You can hit the
       | power button now. I won't get mad; I promise."
       | 
       | That's a somewhat unsatisfactory answer, or more of a software-
       | minded answer. The full answer: Because ATX power was a NEW
       | thing. AT motherboards and power supplies were overwhelmingly
       | still in-use.
       | 
       | https://en.wikipedia.org/wiki/PS-ON_Signal
        
         | Joker_vD wrote:
         | > And back in the day, most computers fell into the second
         | category. Power management was one of those newfangled thingies
         | that only the fancy-pants computers supported.
        
           | 1970-01-01 wrote:
           | PS-ON and ACPI states are different things. You could fully
           | shutdown an ATX system without ACPI 1.0
           | 
           | https://electronics.stackexchange.com/questions/204205/what-.
           | ..
        
       ___________________________________________________________________
       (page generated 2025-02-16 23:00 UTC)