This thread is getting really messy and old. I'll post
here but also make a new post later this afternoon if this
is no longer getting read. BTW, is there some way to get
email notification when a new post appears to a thread of
interest?
Also, Wolf (if you read this) I got your message about
pasting a JPG into my post rather than forcing a browser
to show it. How do you do that? ctl-v won't paste my
file here.
Now, onto the business at hand. I'm still unable to get
W2k to update my driver, or to remove the generic disk.sys
driver that it insists on installing. Wolf is correct
that the "plain vanilla" driver should operate the device,
and it does in my case. It's just that it operates the CD-
DVD/RW/-R portion of the hardware but not the DVD-RAM
part, which shows as a separate device in device manager
when everything is operating properly.
I've used all the tricks that I have to get the updated
driver installed, including just pasting it into my
WINNT/system32/drivers folder. W2k continues to write the
generic driver "disk.sys." on boot. I've also tried
removing disk.inf and disk.pnf from the winnt/inf folder,
thinking that those may be the "writers" of the generic
driver.
I agree that W2k doesn't handle drivers well, and this
specific problem has me thinking about changing my OS.
Does XP do things better? Of course I'm concerned also
that I may buy/install XP and have the identical problem.
Don't want to do that!
Other step-by-step stuff I've done:
- Went through "update Driver" routine
- Windows STILL installs disk.sys driver in Drivers
folder
- If I select "search for a compatible driver
(recommended) I browse to "new drivers location and click
ok
- Wizard returns the inf/disk.inf and installs the
disk.sys driver
- At upate device driver wizard selected 'Display a
list of known drivers for this device so that I can choose
a specific driver"
- After clicking next I get "optical disk drive"
highlighted
- Click "Have disk"
- Browse to folder where I have the Panasonic
updated driver info unzipped
- Select dvdram.inf
- Get a message "the specified location does not
contain information about your hardware"
>-----Original Message-----
>On Sat, 06 Dec 2003 17:53:55 GMT, Leonard Severt [MSFT]
wrote:
>
>>While these are not exactly questions let me address
them.
>>
>>Plug and play has been around since Windows 95 and is
not new in Windows
>>2000. You don't want to have to specify a driver when
you plug in a
>>piece of hardware, you want the best driver to be
installed based on a
>>PNP ID.
>
>Sorry, I disagree. W2K may well have the correct driver,
but it should always
>ask permission, just in case you have an updated driver
for your new hardware
>-- and that is the time to load the updated driver, not
before (as you
>suggest on your next comment below.) In any case, all
peripherals should be
>designed to run in some "plain vanilla" mode, as video
cards all do - they
>will run in plain vanilla VGA if you have no other driver
available.
>
>> If you need to change the driver then change it. If the
3rd
>>party driver has a setup program then it can be
installed before hand
>>and if its PNP ID is closer to the hardware device it
will be the one
>>that is installed. This will not change.
>
>But W2K doesn't behave nicely WRT drivers - it insists on
installing whatever
>it has (and it copies all its drivers from the CD to the
HD, which means it
>often has a driver available, and installs it - that's
what the OP was having
>trouble with!) W2K should it should always ask permission
to install a
>driver, or request one. And uninstalling a driver should
be done cleanly -
>which often doesn't happen.
>
>"Plug'n'Play" is one of the funniest terms in the PC
world, IMO. :-) Wish it
>weren't. :-(
>
>>What do you expect to happen if you change a motherboard?
>
>Nothing whatsoever. Just plug in the HD, and run. Plug in
any peripherals,
>and run. If mobo makers can't come up with common
standards to permit this,
>then they are stoopid. I mean, _really_ stoopid.
>
>> If they are
>>fairly close to same design then great. If they have
different HAL or
>>mass storage controller then you will have problems
>
>That's BIOS issue - the BIOS should present the hardware
in exactly the same
>way to all OSs. There is no need for the OS to know
hardware details. In
>fact, and OS should _not_ know hardware details. That's a
DOS legacy, and a
>stoopid one. -- If hardware makers can't come up with
common standards to
>permit this, then they are stoopid - I mean _really_
stoopid.
>
>> You are changing
>>the base hardware that starts the OS and if changes are
too great the OS
>>can't start because it can't talk to the hardware.
>
>That's a BIOS problem - BIOS should present the hardware
in exactly the same
>way to every OS. There is no need for the OS to know
hardware details. If
>hardware makers can't come up with common standards to
permit this, they are
>stoopid. I mean, _really_ stoopid.
>
>>If by apps you are talking about things like Notepad
then they are very
>>small and any 3rd party program can be installed to take
over the files
>>and options they provide.
>
>Apps, whether bundled with W2K or not, should run from
their home directory,
>with all modules in that directory. (IOW,
_no_ "integration.") The only
>*.dlls etc that they should access outside the home
directory should be the
>common ones supplied by the OS. Each app should have a
sniffer so that it can
>find the OS regardless of what drive or partition it
resides on. (Ideally,
>each app should also have a sniffer so it can find what
OS is present, so
>that it can load the requisite APIs for that OS. All apps
should run on all
>OSs. But that's another rant) Reinstalling an OS, for
whatever reason, should
>have no effect on the apps. An app should run when you
start it, with perhaps
>a few hiccups the first time as it adjusts itself to the
resident (or
>reinstalled) OS. IOW, "installing" an app should consist
of nothing more than
>a) running the sniffer; and b) copying the relevant files.
>
>The real issue is that Microsoft has written its OSs to
do too much. There
>should a clear and clean boundary between hardware, BIOS,
OS, APIs and
>Applications. Microsoft has tangled them up into an
unholy mess - and it's
>the customer that suffers. This NG's plaintive wails for
help almost always
>arise from some stoopid entanglement of these layers. I
mean, when people
>uninstall RealPlayer, and install Nero, why should W2K
suddenly refuse to
>recognise the CD Drive????? GAAAAH!
>
>There's also a security issue here: most of the viruses
work by exploiting
>that entanglement. It's astonishing to me that you can
write a mini-app that
>will affect the way the OS interacts with the BIOS, as
that recent one does
>that causes reboots. How pathetically flawed must an OS
be before people
>decide they've had enough? And its maker decides to
redesign it so it behaves
>as it should?
>
>Ideally: Hardware <---> BIOS <---> any OS <---> APIS <---
> any app.
>
>All an OS should ever send to the BIOS are requests for
certain functions
>(such as "update video display, access printer, transmit
data via external
>port, etc") and data. All it should ever get from BIOS
are requests for
>certain functions (such as "transmit video data, transmit
printer data,
>transmit data for external port, etc") and data. The
only "privileged"
>hardware access for the OS should be the CPU pipeline --
and then only when
>activated by the BIOS after loading the OS. No direct
access to the HD, no
>bother with file systems (those are the HD's
responsibility, after all), etc.
>
>IOW, what we need are standardised data formats and
protocols. Then universal
>computers, capable of running anything the user wants to
run, would be a
>reality. We are halfway there -- the internet uses
standard data formats and
>protocols. Any computer that uses these can access
the 'net. Doing the same
>thing inside the box shouldn't be too difficult. :-)
>
>If you've read this far, you'll see I've come a long way
from commenting on
>Leonard's comments. I'm fed up with the mess. And it's
not just Microsoft's
>fault, BTW. The whole IT community is at fault - there
are too many people
>who "have always done it this way", and can't seem to get
their minds around
>different ways of looking at the box. Even Linux suffers
from legacy concepts
>that have outlived their usefulness, the worst being that
the OS should
>communicate directly with the hardware. Modern hardware
is fast enough that
>the penalty of indirect communication is more than offset
by the advantages
>of a truly smart BIOS. It's time to clean up the mess.
>
>
>--
>Wolf Kirchmeir, Blind River ON Canada
>"Nature does not deal in rewards or punishments, but only
in consequences."
>(Robert Ingersoll)
>
>
>
>.
>
Similar Threads
