It is a couple of things.
"Sometimes" (YMMV), there is a "soft-off" setting in the BIOS.
Mine is a choice of instant or four seconds. (Award V6~)

After the OS is loaded, there are also different things that can,
(or are able to), happen depending on which HAL (hardware abstraction layer)
is installed, and the various ACPI functions that "might be" available in those
different HALs.

These are all power management functions, and MS has had trouble with them
for quite a while. Part of that is caused by themselves since they change the
requirements for the hardware manufactures during the OS design phase.
The hardware manufactures are always playing "catch-up".
XP handles it differently... It intercepts some(?) calls to the BIOS, turns off the
functions itself, and then lies to the BIOS saying everything is fine and the call
was a success. An event log message sample follows:  (Snipped)
    Event Type: Error       Event Source: ACPI
    Event ID: 4      Date:  6/24/02
    Description:
    AMLI: ACPI BIOS is attempting to read from an illegal IO port address (0xcfc),
    which lies in the 0xcf8 - 0xcff protected address range.
    This could lead to system instability.
    Please contact your system vendor for technical assistance.

I'm getting too far from the original question...    Rick Glazier

> I'm guessing There  is a time delay set up into the cmos whereas of you keep
> the power button pressed in 4 sec it will go off.
>
> Fred Caldwell
> Computerboy
>
> > > What does it mean (besides demon possession) when a Gateway won't shut
> > off?
> .  Pressing the on/off button does nothing

         PCBUILD maintains hundreds of useful files for download
                     visit our download web page at:
                  http://freepctech.com/downloads.shtml