Slashdot Mirror

Of course BBC can solve this problem. Other sites solve it. They all agree within one second of my own system's NTP-synced time.

Try man ntpdate. With the right option it will set your clock one time post haste. It's intended for use at startup, when the clock is discontinuous anyway. Then ntpd can keep it synced.

You could do the equivalent of ntpdate when you load the page, and the equivalent of ntpd while the page is displayed, with the clock updating.

Yes, what NTP does is very nontrivial, but it's a long-solved problem. Pulling the NTP levers using the already-compiled executables is not hard, and rolling your own implementation is not exactly rocket science, since the protocol definition is exhaustively defined and open source implementations are freely available.

I'm not going to do the complete design for you, but just consider, why would they use HTTP?

Of course BBC can solve this problem. Other sites solve it. They all agree within one second of my own system's NTP-synced time.

Try man ntpdate. With the right option it will set your clock one time post haste. It's intended for use at startup, when the clock is discontinuous anyway. Then ntpd can keep it synced.

You could do the equivalent of ntpdate when you load the page, and the equivalent of ntpd while the page is displayed, with the clock updating.

Yes, what NTP does is very nontrivial, but it's a long-solved problem. Pulling the NTP levers using the already-compiled executables is not hard, and rolling your own implementation is not exactly rocket science, since the protocol definition is exhaustively defined and open source implementations are freely available.

I'm not going to do the complete design for you, but just consider, why would they use HTTP?

Sigh. The fact that you think British time equals GMT speaks volumes to your lack of understanding of the complexities of time.

(And what is "British time"? Do you mean UK time? What about overseas territories?)

More importantly, if the clock on the user's own computer isn't "good enough", what is? Just agreeing on the requirements could easily take 100 staff days.

The US government has http://www.time.gov/ which has most definitely taken 100 staff days to create, plus on-going maintenance.

There are several things involved:

Firstly, they need to get user's timezone. There are javascript methods to do this, but are not always reliable, especially if they don't want to depend on the client having javascript support. Of course, they could always just ask the user to pick the timezone, so that issue could be solved. E.g. Formula 1 solved it nicely, though I am not sure which method exactly they use (their javascripts are not obfuscated, but I can't be bothered).

Bigger issue, in my opinion, is showing exact time. Assuming their servers all keep exact time and that everybody is happy with their definition of the exact time (which is a big assumption to begin with), BBC would also need to take into account latency between server and client. E.g. it takes about 1/3 or 1/4 of a second for me to load a single random page with a GET request from BBC.

For an example of pain it takes to give users correct time, visit The official U.S. Time page. It's a java applet, presumably because anything client-side can't be trusted to actually count a second as a second. Granted, that page is ancient, but you can still see that it's not really trivial.

The US National Institute of Standards and Technology had something like that not too long ago to display a .gif clock that updated every second for x number of seconds (specified in a URL parameter). They got rid of that, though: now they have either a Java applet or a static (non-updating) gif.

http://www.time.gov/

Take a look at some of what NIST does

Ok - I'll look at http://www.time.gov/timezone.cgi?Central/d/-6/java/

they adjusted the incorrect timestamps from the pictures according to the incorrect time of the camera's clock (not rocket science, I know, but pretty decent deductive reasoning for a cop :-)

I once did that to synch a bunch of pictures taken on two separate cameras. One of the cameras had the wrong date, so they didn't sort by Date-Modified. I wrote a .hta that read the files' Date-Modified attribute, offset it by a defined value, and prefixed the filename with the correct date/timestamp so they sorted alphabetically into chronological order.

Guess how I got the correct time offset? Took a digital picture of time.gov and made the timestamp match the picture... ;)

use www.time.gov

3 Minutes?!?

  I have my machines synced via ntp. ntpq reports than I'm no more than 3ms out of sync with a stratum 1 time server (9ms out of sync with UNSO) and that server is synced with GPS and USNO which as you said is never more than .0001ms out of sync with UTC.

    Eye-balling like you described I can verify that I am within 2000ms of http://time.gov/. I think perhaps that that website may have had issue on the date you saw it being 3 minutes different than what NTP provided.

I'd show you the ntpq output but the lameness filters prevent it.

The problem with these time servers is that many of them are not accurate. I use Ubuntu and when looking through all of the time servers, nearly all of them are off by as much as three minutes. I use http://time.gov/ as my source for accurate time. The atomic clocks are run by the Department of Commerce agency, the National Institute of Standards and Technology (NIST), and its military counterpart, the U. S. Naval Observatory (USNO). Readings from the clocks of these agencies contribute to world time, called Coordinated Universal Time (UTC). The time maintained by both agencies should never differ by more than 0.0000001 seconds from UTC.

How does everyone set their clocks without calling time?

Time.gov.

Official US Government time :

http://time.gov/

Actually, I go to http://www.time.gov/ .

Try this: http://www.time.gov/

I'm surprised that no one has called me on my claim yet, but here's a sample demonstration anyway. Open the following the links in your favorite browser, in the order listed, keeping the IDG site at the top after you're done:
Shockwave site
Java site
Flash site

When I do that, VM Size goes over 150 MB and CPU usage goes over 20% in Firefox, IE, and Opera. It's not up to 250 MB and 50%, but you get the idea with just three tabs and a minute of browser usage. Just add a few more resource-heavy sites, and you can reach those numbers. Add in hours of heavy browsing on a variety of sites, and memory use in any browser can reach hundreds of megabytes.

Now it's time to provide the counter demonstration. Can someone provide a list of links that will cause Firefox to gobble up hideous amounts of memory, but not other browsers? If so, then finally we'll have some details about this problem once and for all!

Just to make my vague, general statements concrete, I picked three sites at random, each of which uses a different plugin:
The official US time clock (Java)
weatcher.com interactive map (Flash)
Panda Pang (Shockwave for Director)

With these three pages open Firefox 2.0.0.3 on Windows XP has a VM Size of 175 MB. Huge memory problem in Firefox? No, Opera 9.10 on Windows XP has a VM Size of 171 MB. After closing the tabs in Firefox, VM Size goes down to 46 MB. Doing the same in Opera, VM Size goes down to 59 MB. If anything, it looks like Opera may have a problem releasing unused memory. Keep in mind for a fair comparison that you must open only those sites after starting the browser, otherwise, you could see the built-up memory usage form hours or days of use in a browser that you've been visiting other pages in.

If you can come up with a series of steps that causes high memory usage in Firefox, and not high memory usage in other browsers, maybe you're on to something.

Check the US time zone. http://www.time.gov/timezone.cgi?Central/d/-6/java

I have an old Macintosh Quadra 660AV, and it came with Norton. Really funny, because Norton would pop up and say, "It seems that your system clock needs adjusting.. Would you like me to fix it?"
Naturally, I always clicked "yes", and was rewarded by Norton.
You know, that is always a good question to ask, because only one clock in the USA is probably correct, and that is the one that the government uses for the official US time.
After Norton fixed my clock, then I really had no way of knowing if it was right or not.
So, after that, I tend to look at all Norton products the same way, a lot of what is done is just for a show, and really does not do anything. Sure takes a long time, however, I have one box that Norton takes 45 minutes to scan all 320 GB for viruses. That's 45 minutes that I could have spent surfing the web on a livecd linux system like mine, Rapidweather Remaster of Knoppix Linux. (Also, see screenshots.)
I am not sorry for the shameless plug, but people will use Windows, and suffer all the faults, and pay whatever Microsoft wants. Now, they can hardly wait for Vista. It'll cost an arm and a leg. But, it has been beta tested by millions, and supposedly, all bugs have been fixed, avoided, or involved in a work-a-round.
We won't know, because the code for all things Vista is hidden. "No idea how they do that." is all that can be said for what you see when you run XP, and soon Vista. Maybe the Chinese will force Microsoft into showing them the Vista source code, so they can make sure the "State Department" does not have a backdoor to Chinese computers running Microsoft products. Didn't they do that with XP?

-- Rapidweather

load this page

http://time.gov/timezone.cgi?Eastern/d/-5/java

now disconnect your network connection...

If this were done (using quasar emissions as a random number source for encryption), what would prevent a third party from surreptitiously generating pseudonoise locally that would mask the quasar emissions? I.e.:

Assume a spy Alice and her controller Bob are both based in the city Xerces, and are using quasar emission encryption [QEE] (I'm copyrighting that acronym-- you have a non-exlcusive license to use it in any way you want). If Xercians broadcast an apparently random stream that drowned out the quasar emissions that Alice and Bob were using, they could break the QEE if they could also learn the method of synchronization that was being used. Even if they were unable to break the QEE, they could assure that QEE messages failed by broadcasting their own "noise" over just Bob's area or just Alice's area (so Alice and Bob would not be receiving the same QEE code).

I do think the idea of having a universally available true random number generator is fantastic! (That would be a QTRNG-- copyright by author, all rights hereby released under non-exclusive license to everybody.) Building a radio receiver tuned to a quasar wouldn't add much to the cost of a PC and having true randomness on tap on desktop machines would open up a lot of new possibilities for Monte Carlo simulations, games, etc. Coupled with access to a third party clock, it would be possible for multiple computers to be using the same random number stream, which could lead to some very interesting things. This is all very good, and I want to see it happen.

(BTW, these methods cannot be patented since as you read this, the above paragraph becomes prior art in the public domain.)

I think that QTRNG could lead to some really interesting applications where the security of the random number stream isn't a big concern, but I don't see how an effective QEE could be developed.

'Even current synthetic benchmarks designed to show off theoretical rates in 3D hardware can have a hard time exploiting the tripling in fragment processing ability.'

If synthetic benchmarks can't show off theoretical rates, how is an actual benchmark with an actual rate going to exploit its processing abilities?

Clock-for-clock all time should be the same.

I run it once every couple of days. Works for me. YMMV.

NTP from a public server is way overkill for most uses. Everyone talks about how easy it is to use - until things start going wrong. There are just too many moving parts that can break.

Actually, if you want to be really specific about it, the Eastern time zone should end about half-way through Ohio. All of Michigan should be in Central time. That's what's causing the most discussion in my neck of the woods (northwestern Indiana), since there seems to be some consensus that since we're near Michigan, we should go to Eastern.

This is the problem with local governments deciding time zones. Everyone has a half-baked reason why we should be eastern, when if you look at the freakin' map, we're clearly central. But, since we want to be on the same time as Detroit, NYC and Boston, we'll end up like Michigan and parts of Kentucky in the Eastern time zone while completely ignoring those pesky scientists screaming about their logic.

Sorry. That's been simmering for a while.

Java's bloat doesn't seem to be that much. I opened up a fresh copy of Safari, and loaded the current time page from time.gov (e.g. for my time zone). Real memory usage was about 13.5MB (about 161 MB virtual). Loading the Java version of that page boosted that to about 30MB (383 MB virtual). What's interesting about that page is that they use Java only for doing the polling of the current time from the server, while doing the display updating (including figuring out the light/dark display on the map) in Javascript.

By comparison, with two windows, each with several tabs, and with the JVM already initialized, I'm using a total of 188MB real memory and 630MB of virtual - and now it's also taking up about 25% of my CPU on this Slashdot response page (either from the text box, or from the animated GIF, I'm not sure), even without doing any typing. Typing into it brings it up another 5%. 5 Freaking Percent just to process my keys and plot some text! How ever did we get anything done on 1MHz 8-bit processors? It would take about 100 of them just to do the processing needed to type some text and word-wrap it. Now it's running at 70% CPU for some reason. Crazy. Running Google Maps, scrolling around or zooming in or out with the satellite view uses about 50% CPU. Just loading the map brings real memory use up above that needed for loading Java (though only increasing virtual by.about 20MB), though that obviously grows as you scroll around and it caches more of the map.

Java's bloat doesn't seem to be that much. I opened up a fresh copy of Safari, and loaded the current time page from time.gov (e.g. for my time zone). Real memory usage was about 13.5MB (about 161 MB virtual). Loading the Java version of that page boosted that to about 30MB (383 MB virtual). What's interesting about that page is that they use Java only for doing the polling of the current time from the server, while doing the display updating (including figuring out the light/dark display on the map) in Javascript.

By comparison, with two windows, each with several tabs, and with the JVM already initialized, I'm using a total of 188MB real memory and 630MB of virtual - and now it's also taking up about 25% of my CPU on this Slashdot response page (either from the text box, or from the animated GIF, I'm not sure), even without doing any typing. Typing into it brings it up another 5%. 5 Freaking Percent just to process my keys and plot some text! How ever did we get anything done on 1MHz 8-bit processors? It would take about 100 of them just to do the processing needed to type some text and word-wrap it. Now it's running at 70% CPU for some reason. Crazy. Running Google Maps, scrolling around or zooming in or out with the satellite view uses about 50% CPU. Just loading the map brings real memory use up above that needed for loading Java (though only increasing virtual by.about 20MB), though that obviously grows as you scroll around and it caches more of the map.

To see the animated clock, be sure that java is enabled on your browser. Otherwise, try a different or newer browser. Also, you may be behind a firewall and you will need to have your network administrator open port 8013 in order for it to work.

It would require that any reader be set with an unhackable internal clock

https://time.gov/ anyone? (No, the link doesn't work ... yet.)

that knows the time zone the reader is in

The author lives in UK, so use UK time (which during the summer is not GMT).

It seems that more and more of everything is sync'd with this. My clock radio at home auto-updates, clock on the wall, the cellphones, my Linux and Mac PC's and cable box.

Only thing left are the clocks with a single AA battery on the wall, and at some point they are going to use the pervasive WWVB time signal that is broadcast from Colorado and operated by the National Institute of Standards and Technology

This technology has really come a long way and is deeply embedded within our lives. Especially if you consider that before the atomic clock, time varied considerably between different locales.

Listen, if we didn't give you useless links, what would you spend your day reading?

Seriously though, how about linking to a less-annoying version of Quicktime?

Guam would be a bad choice. It's on the other side of the International Date line, so it's already tomorrow there. Should procrastinators use the server you suggest, their taxes will be stamped one day later than they would be in the continental US.

On a related note, this is one of the most elegant apps on the Web.

I've been an April 15 filer the last five years. The Post Offices stay open late (at least the bigger ones do), and there's a long line of people and cars dropping off taxes at the last possible minute. It's usually a festive occasion-- ("Hey! We're all filing taxes at the last possible minute!"). It's like the way I've seen New Orleans funerals portrayed. Sad, but happy.

A year is 365.24 days, and its convienient to have New Years at midnight every Dec. 31st, so every 4 years we add a day to account for the time we ignored the previous 3 years. The extensions of the rule exist to account for the fact that its not exactly 1/4 day off.

As for the clocks, the government tells us what time it is (See here and here). Everybody else is pretty much responsible for synchronizing to that time on their own... you adjust your personal timepieces when you start missing TV shows, being late for work, etc. For the most part, its a collective agreement (we had time/timepieces before the US government) and people adjust to an equilibrium because they have to coordinate activities.

If you can run perl scripts on the server, grab http://nist.time.gov/timezone.cgi?/d/0, where the 0 is the timezone offset (-5 for Eastern US time), then parse the time and date out of that. Once you have those values, use the date and clock functions to set your system time.

changing the date on your [PC's clock] fixed [older shareware] pretty handily.

However, newer shareware often downloads ads from the Internet (see Opera) and becomes freeware. Either that, or it goes online and checks for cryptographically signed timestamps that ultimately originate from the official U.S. time.

Does it worry anyone else that it's the Texas flag on the time.gov site?

A handy source of human-readable time (claiming 300-400 msec acccuracy) can be found at www.time.gov. It puts up a java applet with a ticking digital clock. This can be useful when you need to set something manually within a second or two, perhaps for later refinement via a synchronization protocol such as ntp.

Maybe time.gov is better..... ;) check this out

With tiny personal firewall (v2, the free one) my xp box doesn't "call home" every time I search my disk or synch my time. First of all, I don't ever ever let explorer.exe connect to the outside world. Second, I change the time server to time.nist.gov under the "Internet Time" tab of the date/time cpl (check it out at http://nist.time.gov/; yes, I know nist and time are switched, but that's how they have it). Now, if nist.gov was an issue, there might be problems. But last I heard, it was degrees better than "calling home".

I don't know if one can add time servers (perhaps in the registry? never read anything about it), but it would be very nice to find out one could.

Perhaps I'm biased, since I used to work there, but NIST (the government agency that sponsored the study) certainly pumps more into the economy than they take out. This is a Department of Commerce agency, and their mission, in essence, is to support industry in ways that industry can't do by itself.

Partly this means inventing and maintaining standards and techniques for measurement (most famously the atomic clock at time.gov). NIST also developed standard test suites for compilers, and provides lots of free, high-quality scientific software, and so on.

I dunno what everyone is expecting here... but there has been alot of on-line government resources for a long time now.

Consider: St. Paul.gov where you can view this week's prostitution arrests in the city of St. Paul.

What more do you REALLY want from your government than a good laugh?

Do you not own a watch? Here's the Time? Obviously your government is looking after you, considering your every possible need.

As far as I'm concerned these guys need to waste less money on this crap when there are former-tax-payers starving in the streets right now.

Funny I saw an article about this on infoworld last monday... I guess slashdot can't keep up with the pace ;-) I'm sure the article is cached somewhere on google :-O

TastesLikeHerringFlavoredChicken

• 250 GPH Submersible water pump $34 - I think you can get away with a smaller unit, but this runs fairly quiet, especially when put in a cabinet.
• Maze2 Waterblock $42 - This, in my opinion, is the best waterblock out there, and the price ain't bad.
• Transmission oil cooler $44 - dangerden also sells these, since in my exp. carparts.com may take months to deliver.
• Assorted hardware and tubing <$20 - Some 3/8" ID tubing, some hose clamps, and an adapter convert the 1/2" pump to 3/8".
• Some time $priceless - Anyone got some of this for sale? I can use some!

Slap it all together and you're in business. I've removed two of the fans from the case, and put one of them on the radiator but I never turn it on. The one in the power supply runs at low speed until the PSU gets really hot (which it never does now since all the heat is piped out of the case). The hard drives (2x 40GB 7,200RPM IBM 60GXPs) still make some noise, but I put the case on, and put the box in the cabinet in my desk, and I can't even tell if it is running. The ThinkNIC named littlelarry with the fan removed from its heatsink now makes more noise.

Or just go to Time.Gov It's a tid bit easier.

Is this just a localized network problem on my connection, or can anyone else hit it?