Slashdot Mirror
Xdaliclock Fails Y2k (But Everything Else Seems Fine)
Tracy R Reed writes "Like any real geek I was near my computer and used the xntpd-synchronized time to determine when midnight really struck. As soon as it happened, xdaliclock did something strange!" Besides the terrifying xdaliclock crisis, 2600 had a great page up that seemed to fool quite a number of Slashdot readers. Several other joke websites popped up, and several others had real (minor) glitches. So far I've heard rumors of an ATM system that went down for a few minutes, and some radiation monitors that messed up for a bit. But apparently that was about it. The most overhyped event in years. Enjoy the day off if you get one!
Please. Most overhyped event in years? If it were not hyped at or near the level it was, we would have seen disasters. Call it the most useful hype in years.
Kari Hurtta at the Finnish Meteorological Institute has been maintaining a patched version of Elm for a while; here's a page with his patches.
Try filling the gas tank.
---
"'Is not a quine' is not a quine" is a quine.
"'Is not a quine' is not a quine" is a quine.
Quine "quine?
Say the software is only expecting dates from 1980 onwards. Then the Y2k bug hits and gives it a date of 1900. This figure is way out of the ballpark.
If the software is well written, it should see that this is the case and raise an error and maybe stop working. However, this is not always the case and maybe the software performs no check for a valid date - it simply takes whatever is passed to it - in this case, the year 1900.
Now because the date is so wrong to what the software is expecting, it may cause the software to act unpredictably - maybe it might write data to the wrong memory area, cause a stack overflow etc.
Bingo - the whole system malfunctions.
So the Y2k problem is not so much that the system uses dates, but that it also might cause a system with minimal error checking to behave unpredictably.
Field Marshall Stack dun said:
First off, it was a slow busy (literally as if half the connection had dropped); on my end the phone literally hung up, whilst on my dad's side the line was still on (in other words, it thought it still had a connection).
Secondly, I know how Bellsouth's lines tend to behave during busy traffic periods (for example, after the tornado that hit the Mt. Washington area); you will either get a fast busy or you will get a message stating all circuits are busy (I had tried to call my folks to let them know I was ok and making sure they were ok; they live close to Brooks, where the tornado touched down, and I was in the Mount Washington area where the tornado caused F4 damage around a mile northeast of us). Same for when we had a severe snowstorm in 1994 that knocked many of the power and phone lines down (people were asked to stay off the phones, and when the circuits got too busy you either got fast busy signals or "all circuits are busy" messages).
You did not get such funkiness as the phone line hanging up across HALF the connection (which is what it literally did), tying up the phone line to an extent it was impossible to connect to the other side till te circuit finally realised it had hung up on my end. Also, the slow busys were on attempts to RECONNECT; before 7 pm EST we had made a successful call to my husband's folks, and I had begun the call to my folks just before 7 pm EST (which was when the phone hung up and I got slow busy signals).
Furthermore, we had no problems after that point in trying to contact friends whom we were visiting for New Year's, at that.
Furthermore YET, this was at 7 pm EST, not at midnight EST. Nobody reported problems before that time, and I've heard of no problems (save for some stuff with international calls over Bellsouth's circuits, around midnight EST) after that time. It was just the momentary, two-minute hork RIGHT at 7 pm EST (which would be 0000 GMT 1 January 2000); that, and that alone (well, that combined with having witnessed a lot of Bellsouth incompetence and having had to deal with it firsthand to the point that every day I pray to God/Goddess/Cthulhu/[random deity here] that someone else comes in to break Hellsouth's monopoly on phone service--been trying to get bad lines fixed out here for over a YEAR which affect my entire apartment complex AND a major mall, and Bellsouth STILL refuses to admit it's their problem) that made me suspect it could've been a Y2K burp.
I'm prolly gonna check with my sister (who works at a telco) over the next few days to see what sorta funkiness could've caused those symptoms...I've certainly never had the phone lines behave like THAT, though. Not even in Bristol, TN during fall Winston Cup racing weekend (and trust me...it is pretty damned close to impossible to place a call in or out of Bristol on that weekend...even by pay phone, much less cell phones). :P
-Windigo The Feral (NYAR!)
The small West African nation of Gambia has emerged as one of the only countries in the world to be seriously affected by power blackouts and other disruptions in the New Year.
The International Y2K Cooperation Centre, based in the US, says the Gambian energy sector experienced significant power outages while air and sea transportation, the financial sector and government services have been crippled.
Failures have been reported in the Gambian Treasury Department, the National Tax Service and at the Customs Service.
[More at URL...]
Just for spite I decided to turn on my laptop at about 11pm for a couple hours, simply because everyone was saying how "you should turn your computer off until after midnight!" Bah! Linux didn't care. Python didn't care. Emacs didn't care. So pffft! to them :P :)
Ita erat quando hic adveni.
That is not true for all computers. Many older and slower computers use BCD for dates and times. If you are programming an 8-bit microprocessor without hardware multiply/divide, it is easier and more efficient to use BCD. If you need to display or print the date/time, a BCD digit can be converted into an ASCII character with a single addition operation. BCD is also easier for humans who are examining hexadecimal core dumps and packet dumps. Serial time codes such as IRIG-B and NASA-36 are also in BCD format.
Mea navis aericumbens anguillis abundat
This illustrates a real problem in the software world, if you *really* need to know that something is going to work, it takes a tremendous effort to get there, and you're never really 100% sure you've got it right.
This problem isn't going to go away, even if the Y2K (and the 2038?) problems are declared dead issues.
(Sometimes you hear computer science-types speculate about the possibility of automatic software validation, i.e. the possibility of developing a new language where it's possible to mathematically prove that your software is is correct. I'm not holding my breath.)
There are some reports filtering in on the
redhat-talk mailing list about PCs with older
BIOSes that are refusing to boot now.
Don't assume that everything is okay on your
box until after you've rebooted. If you wanted
to be really paranoid, you'd first check the
website of the company that made your mother
board, download any BIOS upgrades, and install
them before reboothing.
(Me, I upgraded my BIOS about a year ago, so I'll
probably just cross my fingers....)
I'm offended too. After working hard to ensure that my sites are Y2K compliant (prevention is better than cure) I now have to contend with a bunch of short-sighted managers saying that Y2K actually turned out to be a waste of resources.
The fact is that if we hadn't made the changes to our systems and infrastructure we would be in a real pickle now.
Oh well.... time to try to educate.
I was just amazed at how badly the BBC did anything. 29.5hrs of continual Millennium coverage
(uninterrupted for anything execpt Eastenders - have to keep those ratings up).
Who really wanted to watch it all?
At least ITV didn't do their usual copycat trick by mirroring everything auntie does.
iain
This is no Y2k bug, it's merely that your wtmp file has been rotated or wiped, so the last login is listed as the Unix Epoch (the date you state is precisely Thu Jan 01, 1970 at midnight UTC, i.e. the Unix Epoch).
It happens to me all the time (because I rotate my wtmp files faster than I should, I guess). Except that since I'm in the +0100 time zone, the date printed is more immediately recognizable as the Unix Epoch.
Well, all my messages going out via elm (my primary mailreader) are sent from the year 100. Great. I've got the word "fool" stamped on all my messages. And, there hasnt been an update for elm in ages. Is this thing an orphan?
Guess I have to dig into the code to save face.
I really appreciate it! I wish I could mod you up!
I thought about that response, and since it's a geek's primary responisbility to educate others (thus boosting his/her own self image), I didn't like my answer...
The 8080 used 8 bit words. The 8088 used 8 bit external bus and internally was 16 bit. Thus any interface it had with other hardware was 8 bit.
A lot of PLC equipment is still based on this standard. Thus the hardware (Input/Ouput cards) communicate with the main processor on an 8 bit backplane.
8 bits give you a maximum integer value of 255. Therefore a year like 1987 won't fit. The best you can do is save the last two digits (ie. 87).
The only fix to this problem was changing the hardware... no software fix available.
Make Sense?
(I guess I didn't want you to do the math after all.)
Even though you could see the digits at the base of the ball spell out "2000" before they lit up, I was desperately hoping that when midnight hit, they would actually light up "1900" instead.
/that/ joke.
/Everybody/ would get
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
(Sometimes you hear computer science-types speculate about the possibility of automatic software validation, i.e. the possibility of developing a new language where it's possible to mathematically prove that your software is is correct. I'm not holding my breath.)
Me neither. Consider:
Suppose you have an automatic software validation system. It will require as input a specification of the correct behavoir of the software, in a formal language that it can parse. Rendering the specification into the formal language amounts to writing a program. How do you prove THAT is correct?
Therefore a general formal correctness prover is impossible. You micht be able to automatically prove that two expressions of a program are equivalent, or that a particular program matches a particular formal specification. But you can't prove that the program is "correct".
Which is not to say that such tools are useless. I have yet to see a method of testing a program that doesn't amount to expressing the program and/or its requirements in two languages - as distinct as possible - and then comparing the two. Typically this will be the code versus a human-readable spec, pseudocode, or profuse comments, and the comparison will be done by humans: a team of programmers or software QA engineers. The process might include code walkthroughs, writing and running automated tests, coverage analysis, etc. But an automated tool comparing code against a formal spec would be a very useful labor-saver, or an additional tool in the kit.
The important thing is that expressing the program in two distinct languages results in two distinct sorts of thinking as the two expressions are written. Most bugs tend to occur only in one of the thinking modes, and are thus exposed when the diverse expressions of the program are compared.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
So Sun Dec 26 went down to Fry's with the wife and got a copy of Red Hat 6.1 and his-and-hers Compaq 5888s. (Didn't get monitors - had some old ones and am expecting a big price break on LCDs this spring.)
Loaded Linux onto mine on Mon and started configuring on Tue. Had to patch the ethernet driver to suport the diamond "homenet" / ethernet card. (Stock Red Hat driver is hardcoded to use it in the "homenet" mode - 10-base-T port dead and trying to do a 1Mbps link on the telephone wiring.) Winmodem is dead of course, I don't think the install found the DVD drive (though the CD/R is working), and the power button is software driven so I have to unplug it to shut it down until I get that configured, but it's close enough for now.
Reason for the migration: I'd been running my home network on an old Sparc with SunOS 4.1.3, and didn't want to drop mail on the floor. Hung an external modem on the serial port. Beat my head against sendmail.cf for a while. Taylor UUCP's automatic config file translator worked beautifully. Finally got it all working and cut over at 9:10 PM New Year's Eve. (Had also copied all the files from the SunOS to the new machine for safekeeping and salvage. Old disk was 2G and hitting the wall. Whole thing takes up Shut everything down (to guard against crossing-the-boundary bugs) and went out to party.
1/1/2000 turned things back on again - and discovered that shutting down had been the last straw for the old Mitac monitor. B-( Horizontal oscilator won't even start with the video in 640x480 mode. (Played musical monitors to check that the computer was OK.)
So back to Fry's and dropped another $400 plus tax on a new 19" Princeton Ultra95e. (NICE, SHARP display. Time to get new glasses so I can take advantage of it.)
The most annoying part, of course, is that SunOS 4.1.3 came up fine and said the date was Jan 1 2000. So it probably would have kept working well enough to handle the mail after all.
Oh, well. At least it gave me an excuse to cut the whole home network over to an Open Source OS and my workstation to a 700 Mhz Athlon.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
(Not an attack- I'm genuinely curious:)
Even on small systems like that, how could Y2K be a problem? Computers store numbers in binary, not decimal, and with the number of bits required to store two BCD digits (that is, 8) you could also store any number between 0 and 255 in a straight binary representation, which would put the Y2K threat on New Year's Day 2156. So why is it that the problem would have happened yesterday?
-jacob
Well spotted - had to look again to see it - I take it that you *really* do read the comics, not just skim through them like some of us :)
I left my body to science, but I'm afraid they've turned it down...
To the people who say "See, Y2K was all bullshit!" :
:)
:)
If you're thinking Y2K was all hype and a load of crap, etc, then you obviously had nothing to do with it or only experienced your own small part. As usual, in these kind of cases, extrapolation to the "whole world" from your small experience is incorrect!
Over the past two years, I worked at the corporate level for a major telecoms company, a state government transport department (trains, buses, etc), a major insurer, a union and a medical education institute. I also advised numerous clients and even wound up digging into software I'd written "on the side." (YUK!
All I can say is that had we done nothing, the phones would have slowly ground to a halt, billing would have been screwed, premiums incorrect, etc etc. Y2K was real and it was only the efforts put in that have resulted in the "anticlimax" that we have experienced to date.
Don't forget, it's still early days yet. Small to medium business hasn't returned from holidays, the Leap Year is still to be passed, end of Financial Year calculations, etc etc etc. The show isn't over and won't be for some time. We still also have the Unix/C epoch date (2038 - yes?) and fixes for all those "windowed" solutions (eg: = 30 = 1999) - what happens in 20 years when that code is still in use?
'Scuse the ranting - just that it burns me a bit to see all the "Told ya so's!"...
I left my body to science, but I'm afraid they've turned it down...
that behaviour in xdaliclock looks more like a planted easteregg than an actual bug... the behaviour really is bizarre. i doubt it's a malfunction.
on that note, did anyone find any "strange" behaviour in other programs? i was expecting windows to do something strange, but apart from crashing as usual, it seems normal. anyone find any other y2k easter eggs in other programs?
Fross
My parent's 486's BIOS has a problem with Y2K. I can set the year in the BIOS, via DOS prompt, or via Windows and it will report the date as 2094 upon booting. Oh well, I'll write a ASM program that will run on bootup...
That's an interesting observation.
How is one supose to tell, exactly, if the one-a-day-BSOD is y2k related or not?
For what it's worth, I got a phone call from my uncle who left a message on my machine at a little past midnight last night, telling me that his windows box 'has that blue screen again' and asked if he should wait till the next day to reboot.
_________________________
Three cable TV systems in Alabama failed at the stroke of midnight.
The wait for tech support doubles every 18 months... Any likelihood they can solve your problem halves. Foosters
I was so sick of people saying the definitely nothing would go wrong and everything would be fine. Sure in the end that was the case - but that is with hindsight and after all the effort that was put in beforehand.
Looking at the posts about various systems still up and running, many of them seem to be of the sort: "Hey, Windows/Linux/whatever is still running on my desktop - bah, I knew that the Y2k bug was all hype."
Having worked in both the engineering and enterprise IT industries, I have worked with many many large systems, and you are right redtoade, many systems are old and archaic ... and also essential in the day to day operation of the business.
Collegues who have worked with me in the past year on the Y2k problem anticipated that the transition would be a fizzer (ie quiet) after all the work that they had done, but no-one would be so naive to claim that Y2k was all hype.
These comments come from guys on the front line. Most people don't have this experience and extrapolate their limited experience to the rest of the world - ie. it was easy to check my system, the rest of the world will be fine.
As redtoade pointed out, many systems failed during testing and it is only through all the preventative efforts that there were no problems on the night.
How anyone could certain that nothing would go wrong is beyond me. Sure, one could be quietly confident (especially after all the effort that was put in), but to be absolutely certain?!? I still shake my head at the number of people who think that the Y2k bug was all scam and groundless hype.
Well, it seems the worst that happened with me and Y2K (besides the 19100 bug on several sites--right now that seems to be the most common bug, which is probably a Good Thing) was a rather odd little hiccup that occured around midnight GMT (I'm in Eastern time, btw)...
I'm calling my folks to let them know I'm off and to wish them early Happy New Year's and all that so they don't get all worried about Y2K and all.
Upon which--mysteriously--the phone hangs up. I try redialing for a good five minutes. Phone line is busy (really damn weird...because my folks have call waiting, and the LAST time the phone line gave busies there was when the tornado hit in the Brooks/Mt. Washington area just south of Louisville in 1995).
I finally get through..."Dad, did you hang up?" "No...I thought YOU hung up..." Both of us are counting this down to a momentary fart on the part of Bellsouth Kentucky, aka the one phone company in English-speaking North America that makes US Worst actually look good. :)
"Bellsouth reporting no problems", my arse. :) At least in Kentucky, I'm fairly sure the phone lines did hiccup...then again, the phone lines are generally wonky here... :)
-Windigo The Feral (NYAR!)
I bet this made a few people's hearts beat faster when it happened.
I'm with you ... I didn't have to do any ASM recoding this year, but I have to agree with your stance. Many people have asked me if Y2K was over hyped. I usually say 'yes and no' -- that is, it wasn't hyped enough to get businesses to react as quickly as they should have (upper management) but it was hyped too much as an end of the world scenario (bunkers and food stores and the like).
- Michael T. Babcock (Yes, I blog)
Did anyone else notice the copyright date? My only complaint about the rot13 is that there is no tool I know from automatically decrypting a comic strip. I had to type it into Emacs and M-x toggle-rot13-mode. Please, nobody make any nasty comments about the fact that I obviously have no aversion to typing.
The net will not be what we demand, but what we make it. Build it well.
Last login: Wed Dec 31 1969 19:00:00 -0500
Unfortunately, you can't fool uptime:
12:32pm up 46 days, 21:03, 1 user, load average: 0.00, 0.00, 0.00
I was stuck at work during the rollover, and the only real problem we had was a fax server thinking that it was in the year 1899. And the countdown screensaver we've had for more than two years finally crashed.
Y2K World Dispatches
Y2K Security Tracker
One of my favorites from the World Dispatches was:
---------
Question: How do I leverage the power of the internet?
---------
There is no try at jedinite.com
For those of us that had to work Dec 31, y2k is real.
In the sense that:
All phones now forward to bosses office.
For somewierd reason random computers had thier bios changed to longer recognize the hard drive.
You now have the best chair in the office.
Various settings on computers are screwed up.
You have uninstalled solitare on bosses computer (going to enjoy him debating wether to ask tech to reinstall just so he play game and look like a fool or just sit there)
Quake III is now what loads if you click on any shortcut (in WinNT)
The clocks are all an hour slow.
People have random meetings scheduled in there planners that they left on the desks.
Backgrounds on monitors show the boss and his secretery... involved in a rare act that his wife would kill to find out.
So bosses out there... the lowly computer programmers you made work Dec 31 - New Years Day... Have gotten thier revenge... muahahhahahhaha
(currently figuring out to change building security code)
Is it progress if a cannibal uses a fork?
Comment removed based on user account deletion
Internal Server Error
The date specified (01-01-1900) is impossible. If you have forced this
error condition, you may be in violation of state, federal, and/or civil
laws. Those outside the United States should check with their respective
governments concerning their country's extradition treaty. Dissemination
of this error is also strictly prohibited.
If you believe you have received this message in error, please reload
the page and try again.
Pretty odd, eh?
--Mike--
With false american pride I ask . .
Here we go:
From the Auckland International Airport Limited we have this little date jem:
Y2K Update
Dated 02:58 1 Jan 100
Auckland Airport Confirms Business as Usual
From the bury-your-head-in-the-sand-till-it-goes-away dept., we have this one from audiusa.com
Sorry, we have temporarily disabled this module. "While Audiusa.com is fully prepared for Y2K and beyond, we wanted to keep our databases clear of the millennium madness everyone seems to be talking about.
From www.tvtoday.de/tv we have this little assending date snafu jem:
Samstag, 01.01.100
Then from the people at pleaseread.com we have another assending date from hell: ;)
Saturday, January 01, 19100
Award-winning text-to-speech software applications utilizing
the best technologies in the world.
(Editors note: Best technology? they can't count to 99
And from the make-it-stop-make-it-stop dept.
we find AllAdvantage.com thinks the best way to comply is to not partisipate:
Happy Y2K from AllAdvantage.com!
As a precautionary measure, we've disconnected our servers from the Internet and are watching the millennial date change from the sidelines.
Most of these small errors (19100) are caused by one small piece (printf("19%d", tm->tm_year);) of sloppy code (as previously posted on /.). While these little didbits make for a good laugh, I'm happy to see that non of them add up to the 'show stopper' everyone has been hyping. Happy New Year!
_________________________
Anyone else notice the countdown error during the ball drop in times Square? As they started counting down minutes and seconds, it would display 15:01, 14:00, 14:59. It did this all the way to midnight. And Sam Donaldson said nothing was wrong.
I'm kind of offended at the tons of posts saying:
"NOTHING HAPPENED!!! See, I told you so."
I'm sure all the IT and CompSCI people out there weren't really worried about any Y2K stuff... but we control systems engineers (level 1 production types) were terrified of it.
Anxious as we were, we tried to get management to fund an inspection of all systems... and since management is usually made up of non-geek types, there was no way to squeeze a penny out of them for what they felt was unneeded computer work.
BUT, thank God for all of the Y2K hype!!! If management hadn't seen Dan Rather explaining (in small terms) what the possibilities for Y2K were, we wouldn't have received time and funding to inspect our systems.
So after 7 or 8 of my clients found 85% of their systems to NOT function after a clock change to Y2K, we spent MILLIONS of dollars this year to fix them PRIOR to the actual event. These weren't Intel based processors mind you... these were Programmable Logic Controllers (PLCs) and older Distributed Control Systems (DCSs), some even based on Z80 and 8080 technology. The industrial/manufacturing world is still decades behind in it's control systems... so while you guys are lounging in your linux alphas... we're still doing machine code.
I guess what I'm saying is Y2K was VERY REAL for us... and you can thank the thousands of engineers, technicians and programmers who fixed the problems BEFORE THEY HAPPENED for a very quiet new year's eve.
Happy New Year everyone.