All OSes need a usage meter!

[foxkm]

Ever looked at piece of machinery and seen a counter on the engine to know how many hours of use the machine
has got. They do this for a few reasons..

1. To keep track of how long the fluids need to be changed
2. To determine the value of the equpitment with depreciation.
3. To figure out how much wear and tear the components have had... etc

Now what if an OS were to have a kernel level module that would monitor and store this same information in a
locked location?

What would the advantage be:

1: If the OS is broken in some way, you can figure how many "miles" the OS has on it and determine if a reinstall
is the best option, or if its worth fiddling around in the os to attempt to fix.

2. Determine how much use in general the hardware has if you are looking for hardware problems.

Anyone have any comments or know if there is a solution already for this purpose?

I know that you can usuall tell when Windows was installed by the creation date of C:\Windows

foxkm

 

[rjain]

digital logic doesn't degrade gracefully unless it has ECC. Even then, that only applies to busses and data storage, not to actual processing circuits. If you do 1+1 and get 32767 because of a single blown transistor, your app will probably not to do the right thing by any stretch of the imagination.

 

[dnuggett]

If the OS is broken in some way, you can figure how many "miles" the OS has on it and determine if a reinstall

An OS isn't a car. The amount of "miles" it has on it has no bearing on if you should reinstall it. The changes that have been made to it are what's relevent not the amount of usage.

 

[Nothinman]

Software doesn't wear out as things with moving parts do, although some versions of Windows do seem to wear out after time for some uknown reason.

Things that do have moving parts, like hard disks, do have counters in them. atleast any drive in the past 8 or so years that supports S.M.A.R.T..

 

[rjain]

Also, recent CPUs have some sort of health monitor that can indicate if it's operating at marginal conditions, which may indicate either worn out circuits or overclocking/overheating

 

[zephyrprime]

Computers aren't like mechanical devices. They don't wear in the same way so this idea is useless.

Things that do have moving parts, like hard disks, do have counters in them. atleast any drive in the past 8 or so years that supports S.M.A.R.T..

I have yet to hear of anyone on anandtech that has had their ass saved by SMART. It seems like when HDs die, they just up and die with no warning.

 

[Nothinman]

I have yet to hear of anyone on anandtech that has had their ass saved by SMART

I've had dozens of drives at work get replaced before they totally failed because of SMART, of course not all of them give a warning but you can't say it's worthless.

 

[rjain]

I've had my ass saved by SMART. Well, if I had acted faster, I wouldn't have lost anything, but I only lost about half my mp3s while waiting for me to order my new drive and for my new drive to come.

 

[drag]

Never heard of SMART before... How you go about using it?

edit: nevermind, did a search for "smart" in dselect.

 

[rjain]

drag, you might need to enable SMART in your BIOS, too.

 

[InlineFive]

Originally posted by: dnuggett

If the OS is broken in some way, you can figure how many "miles" the OS has on it and determine if a reinstall

An OS isn't a car. The amount of "miles" it has on it has no bearing on if you should reinstall it. The changes that have been made to it are what's relevent not the amount of usage.

Well in Windows 9x (95, 98, 98SE, ME) this could be taken as true. After a while you need to reinstall if you use one of those.

However in Windows 2000, XP or 2003 this shouldn't be necessary.

 

[spyordie007]

uptime.exe will give you a report of the total time the system has been up (at least back to the beginning of the event log).