System Clock loses time - help

[fish7892]

My system clock fails to keep time when the Marine Aquarium 2.6 screensaver is on for a while. It loses a few minutes every hour. How can I stop this system clock problem? I have to add some minutes to my clock every morning when I get to work.

This occurs on my Windows 2000 Server machine, a DELL Dimension V400, w/ S3 Graphics ProSavageDDR video...

[Wizwad]

I'd say, offhand, that you need to replace your BIOS battery. Either that, or download one of those programs which keeps your system clock synched to an atomic clock somewhere.

[feldon34]

This is not the firsttime this has been reported.

[FishyBusiness]

Are you saying that the aquarium does indeed affect the system clock???? If so, that's a MAJOR flaw.

[feldon34]

It came up several times during the Marine Aquarium Time betas but I thought it had been resolved. Marine Aquarium Time makes extensive calls to the time functions in Windows. Because Marine Aquarium 2.6 contains the Marine Aquarium Time code, it also makes extensive calls to the time functions.

Why *reading* the clock can actually affect the time is beyond me.

[Jim Sachs]

I do remember someone asking about this before, but I thought that Edgar had looked into it. Like Morgan, I personally don't see how observing the time can affect the clock. Maybe there's a deep Physics issue - the Heisenberg Uncertanty Principle - at work here I'll look at my watch 1000 times today and see if it slows down.

[Surferminn]

Actually, it sounds like the ol' Quantum Physics Double Slit Experiment wherein the observer actually affects the observed event causing it to perform differently than it may when not observed.

[FishyBusiness]

Originally posted by Jim Sachs:

...Like Morgan, I personally don't see how observing the time can affect the clock...  

That's exactly why I posted my previous message.

[Jim Sachs]

I have run the Aquarium on many computers for weeks at a time, and none have ever lost a second of time. But I'm not totally discounting the report - about once a year it seems to come up. It's possible that the clock battery on those motherboards are just winding down.

[Edgar]

Here is a link that Microsoft knows about the time bug on some computers.
http://support.microsoft.com/?kbid=821893

Edit: From the article:
"Modify the program to call the timeBeginPeriod function at startup and to call the timeEndPeriod function on exit. This workaround eliminates repeated time increment changes."

Note the screensaver code does this already.

[jroseone]

Mine keeps time better than my Rolex

[fish7892]

Well, it is losing time. Serene Support has suggested a newer video driver, because mine is a bit out of date. I'll try it. Someone suggested syncing with a time service. Sure - I can do that, but that's not addressing the actual problem.

As someone else suggested, it could be the act of checking time that affects time itself. So, I sat at my desk watching my screensaver all day, and checking my watch every 10 seconds, for 8 hours. And, indeed, my watch sped up! According to my system clock, my watch was 7 minutes faster!

[Jim Sachs]

Ah, that's the problem. You are in a Quantum Time Warp. The computer clock is not slow, your own time has sped up. If the effect develops into a full-fledged wormhole, use it to go into the future. A year or two should be enough for us to find the solution.

[fish7892]

Okay, a real experiment this time. Last night I re-enabled the crystal aquarium clock. I had it off before, because I like the nice and natural look of the aquarium (excellent application, if I haven't said that yet!)

Anyway, the system clock did NOT lose time when I arrived at work this morning. I find that very interesting...

Aquarium clock enabled -> System clock keeps time
Aquarium clock disabled -> System clock loses several minutes per day

At least with my current configuration. Seems like perhaps this could be a useful clue to help resolve my issue...

I suppose I can leave the clock enabled if need be...I just like the aquarium in it's nice & natural state...

[Jim Sachs]

Thanks for the experiment - the results may be useful to Edgar.

[Edgar]

That may be useful. I can at least look in the code to see what calls may be different that may affect the time. Of course if it the time doesn't slow down on my PCs, it will be basically guessing. I might have to ask you later to test it for me.