The Daily WTF: Curious Perversions in Information Technology

chubertdev · 2013-07-24 Reply Admin

hank:
eViLegion:
Why bother getting offended by stereotypes? By not bothering to get offended, life is easier, and for some reason things seem to taste better.
Right or wrong, somebody's always getting offended, and consequently nothing's ever easy.

is that the way things were done back then?

2013-07-24 Reply Admin

You whippersnappers don't even know the algorithm for drawing a circle on a x,y grid using sine and cosine, do you? that's some seriously intense computation there. Modern computers only can do it using approximations of circles and bitmap images. Otherwise your precious 'windows' operating system would come crashing around your knees too.

2013-07-24 Reply Admin

"Do not disturb my circles!" Archimedes last words.

chubertdev · 2013-07-24 Reply Admin

Morry:
You whippersnappers don't even know the algorithm for drawing a circle on a x,y grid using sine and cosine, do you? that's some seriously intense computation there. Modern computers only can do it using approximations of circles and bitmap images. Otherwise your precious 'windows' operating system would come crashing around your knees too.

Stop being so polarizing.

2013-07-24 Reply Admin

Maybe Miklos couldn't read the instructions either because he didn't speak Czech? Miklos (Miklós) is a Hungarian name.

da Doctah · 2013-07-24 Reply Admin

¯\(°_o)/¯ I DUNNO LOL:
Ulrich made sure to add "DO NOT USE CIRCLES" to his Czechlist.

Now if someone will just be so kind as to add it to the tag cloud....

2013-07-24 Reply Admin

Just BTW, we don't know how heavily loaded the mainframe was before Ulrich started this program. Maybe it was already running at 99% of capacity and he just pushed it over the edge. Also, we don't know how many circles he tried to draw simultaneously.

2013-07-24 Reply Admin

One thing that puzzles me about this story: If they didn't have a test system, where did Ulrich develop this code? Did he do his development on the mainframe? I mean, was his demo for all the hot shots the very first time he tried to run the program? I can't imagine giving a demo to anyone before I'd run tests myself and was sure that it worked. How often does a non-trivial program run correctly the first time you try it? (Answer: almost never.) If he ran those tests on the mainframe, wouldn't it have crashed long before the demo? And if he didn't do his development on the mainframe, then why didn't he do the demo on his development system?

Either there's a missing piece here or something has been lost in anonymization.

2013-07-24 Reply Admin

I also wonder why he couldn't just re-start the mainframe and not run his new program. He was apparently able to update his code to remove the circle-drawing and restart. So how did he update the code without the computer running?

2013-07-24 Reply Admin

sqlblindman:
"Do not disturb my circles!" Archimedes last words.

Exactly. That's why circles are bad and you should never use them. If he hadn't drawn circles he would probably still be alive by now.

2013-07-24 Reply Admin

I bet you could pick up one of those mainframes for a song and start up your on plant.

chubertdev · 2013-07-24 Reply Admin

I like how people are trying to guess the location of the story based on fictitious names.

Nagesh · 2013-07-24 Reply Admin

what comic-book is this from? " black as the cloak on a member of the Night’s Watch"

SpewinCoffee · 2013-07-24 Reply Admin

This explains why TDWTF doesn't use complete circles, rounded corners, or anything else round anywhere on the site.

chubertdev · 2013-07-24 Reply Admin

Nagesh:
what comic-book is this from? " black as the cloak on a member of the Night’s Watch"

A Song of Fire and Ice

Better known as Game of Thones

2013-07-24 Reply Admin

In Mother Russia, you can draw circle ... as long as circle have 4 corners.

2013-07-24 Reply Admin

Why couldn't they just roll it back? Unless they were using the mainframe as a development machine in which case we are all doomed. Or did they not have a backup????

2013-07-24 Reply Admin

MrOli:
I have no sympathy for this story...

Until your kids are eating their cereal through their foreheads and have eleventeen toes on each foot.

2013-07-24 Reply Admin

If he was working at the nuclear power plant in the picture, he should already know Czech; that's the Dukovany Nuclear Power Station.

2013-07-24 Reply Admin

DCRoss:
Right or wrong, it's just the way that circles gave always been drawn in nuclear power plants.

Right or wrong, that's the meme we'll have to endure the next two years, brillant ...

2013-07-24 Reply Admin

chubertdev:
Morry:
You whippersnappers don't even know the algorithm for drawing a circle on a x,y grid using sine and cosine, do you? that's some seriously intense computation there. Modern computers only can do it using approximations of circles and bitmap images. Otherwise your precious 'windows' operating system would come crashing around your knees too.

Stop being so polarizing.

You asked for it.

2013-07-24 Reply Admin

jay:
One thing that puzzles me about this story: If they didn't have a test system, where did Ulrich develop this code? Did he do his development on the mainframe? I mean, was his demo for all the hot shots the very first time he tried to run the program? I can't imagine giving a demo to anyone before I'd run tests myself and was sure that it worked. How often does a non-trivial program run correctly the first time you try it? (Answer: almost never.) If he ran those tests on the mainframe, wouldn't it have crashed long before the demo? And if he didn't do his development on the mainframe, then why didn't he do the demo on his development system?
Either there's a missing piece here or something has been lost in anonymization.

The answer is that he tested his changes the same way our editors test their fictionalization for consistency.

(Yes, I know, on a meta-level this statement doesn't make sense.)

2013-07-24 Reply Admin

Bob:
You do a comment and funny not come? You want Miklos to fix this? You stupid? If you know comment not funny, then DO NOT COMMENT!

good advice, I wish you;d take it.

2013-07-24 Reply Admin

I think the mainframe couldn't handle circular logic.

2013-07-24 Reply Admin

Golden Dragon:
Nix Nada:
my name is missing:
I read the words of Miklos in the voice of Peggy.
I read them in the voice of Miss Piggy. "Are you stupid? You want MOI to fix it?"

I read them in the voice of Roman "Niko! Cousin! Stop drawing circles and let's go bowling."

I read silently so I don't disturb others.

2013-07-24 Reply Admin

billy le kid:
Bob:
You do a comment and funny not come? You want Miklos to fix this? You stupid? If you know comment not funny, then DO NOT COMMENT!
good advice, I wish you;d take it.

I thought it was funny!

Coyne · 2013-07-24 Reply Admin

Patient: "Doc! It hurts when I raise my arm like this..."

Doctor, interrupting and yanking patient's arm down, "Then don't do that!"

2013-07-24 Reply Admin

i do not believe this story for even one minute, sorry

2013-07-24 Reply Admin

Not a U.S. nuclear power plant. They have simulators (identical mock ups of the control rooms) that are used to train operators that could be used for a test environment.

2013-07-24 Reply Admin

no laughing matter:
Shoreline:
I'm pretty sure this is TRWTF:
They didn’t have a “test control room”...
This is a nuclear power plant, and they can't test what happens in an equivalent environment. The plant will blow up. We are all fucked.
The cautious guys at Fukushima had not only one, but three reserve plants to test what happens when you operate a nuclear reactor in a region prone to earthquakes and tsunamis. The test results showed that the plant will blow up, which could be reliably reproduced in subsequent tests!

No they did not test real tsunami. They might have tested earthquakes of magnitudes that had occured in the past; I don't know if they did. But they didn't test tsunami of heights that had occurred in 1896 and 1933. It was beyond their imagination that there could ever be a tsunami as high as those that occurred in 1896 and 1933.

2013-07-24 Reply Admin

xaade:
I'd understand if the display failed, but drawing circles crashed the mainframe? I mean, did the software cause the equivalent of BSoD? The whole OS went down.

The crash might have been the equivalent of a BSOD or it might have just been a hang.

1970s: On IBM mainframes, if a program aborted (abended) and had not used up its limit on CPU time, the OS continued charging CPU time against the program while producing a hex dump. If the program's CPU time limit was reached while producing the dump, the OS abended the program again. The OS couldn't handle its own recursive abend of a program, and the result was equivalent to a BSOD.

2010s: On a PC with an AMD (ATI) graphics chip and Windows 7, if AMD's driver hangs then Windows 7 restarts the driver -- once. If AMD's driver hangs a second time then Windows 7 hangs. No BSOD, just a hang. The only way to restart is a power cycle, losing your work the same as a BSOD would do.

2013-07-24 Reply Admin

Methusalem:
There's probably a real-time driver running on the same mainframe, polling sensors, or updating the displays or something. And no modern operating system capable of enforcing hard real-time constraints. When Ulrich drew circles in some painfully slow trigonometric-math fashion, the timing on some driver went off. If that driver was for the display, then screens going black is a likely result.
That's the 1970's for you.

That's also the 2010's for you. Meanwhile, in the 1970's some vendors other than IBM could make hard real-time operating systems, and in the 2010's some vendors other than Microsoft can make hard real-time operating systems.

2013-07-24 Reply Admin

¯\(°_o)/¯ I DUNNO LOL:
And I remember working on a system in the mid 80s that had all the requirements for hard real-time operation, notably the fact that if high-priority tasks are running, low priority tasks don't get a time slice, *ever*.
I used VAX/VMS back then too, when I was in college.

VMS did not work that way. When a low priority task had a page fault, its priority got bumped 6 levels.

Start a task at priority 1, enormously low. Start another task at priority 6, moderate. The moderate level task does I/O or something so the low priority task gets a CPU slice. It hits a page fault and waits its turn, priority now 12. VMS bumps its priority from 1 to 7. The moderate priority task gets its I/O finished and resumes running at priority 12 but quickly drops back to priority 6. The low priority task gets its page swap finished and takes the CPU because it has priority 7 but quickly drops back to priority 1. So the low priority task doesn't get much CPU, the low priority task takes forever to run because it's low priority, BUT IT IMPACTS THE MEDIUM PRIORITY TASK AND SLOWS THAT ONE DOWN BY A FACTOR OF 2.

32 levels had 16 levels of real time stuff and 16 levels of ordinary stuff. It wasn't enough. If we wanted 5 levels of priorities, and we didn't want high priority tasks to be impacted by the page faults of low priority tasks, we had to separate our 5 levels by at least 7 VMS levels between each of our levels. We needed around 29 levels of ordinary stuff and 29 levels of real time stuff. We made do with 3 levels of ordinary stuff. Performance for interactive users almost became bearable.

A few years later I got a job at Digital, but by that time DEC was dead and already taken over by Digital. The days were long gone when an engineer could discuss stuff like that with anyone that had rank to fix anything like that.

2013-07-24 Reply Admin

I'm a WTF today, posting two responses impersonating someone who should have just been impersonated one time yesterday. Anyway...

belzebub:
* Old computers didn't have any graphics acceleration, and often no graphics libraries. There could've been some form of "ClearScreen", "DrawPixel" if you were lucky.

So, when a programmer in Czech Republic was working on mainframe in 80's (they was not manufactured after the fall of communism), and he had to write circle drawing function, he would probably do something like [snip]

Also, as I'm Czech, this is super offensive :)

You should have copied a machine from TRW instead of DEC. They handled circles in the 1970's.

lolwtf · 2013-07-24 Reply Admin

xaade:
Dr. Azrael Tod:
so this was 1999 and they used old mainframes from the Czech Republic build in the 1970s. I would say it should be safe to assume that since those mainframes were older than 10 years and thus lived through the fall of the eastern block.
Who ever wrote that "WTF-Article" didn't quite get how far away from concerned this producer of toasters is from caring about bugs in archaic computers. It's pretty improbable that they have anything in common with the organisation that built it besides the pure name.

Whereas I agree that Miklos shouldn't be expected to fix software that he likely doesn't have the source-code for, the WTF is that circles did in the entire machine. I'd understand if the display failed, but drawing circles crashed the mainframe? I mean, did the software cause the equivalent of BSoD? The whole OS went down.

Now, I understand Miklos' point of view. Imagine someone calling up Microsoft:

"Yeah, uh, Windows 3.1 crashes if I try to use API ____."

"Seriously? You want me to fix 3.1? I'm sorry, we don't even have a backup of the source code for 3.1, much less a machine that will run it. If API ____ fails, then don't use it."

And to anyone who's used Windows' graphic APIs, and especially Win3.1, the idea of drawing a circle causing the whole OS to crash doesn't sound at all farfetched.

da Doctah · 2013-07-25 Reply Admin

Tracy Kidder:
no laughing matter:
The cautious guys at Fukushima had not only one, but three reserve plants to test what happens when you operate a nuclear reactor in a region prone to earthquakes and tsunamis. The test results showed that the plant will blow up, which could be reliably reproduced in subsequent tests!
No they did not test real tsunami. They might have tested earthquakes of magnitudes that had occured in the past; I don't know if they did. But they didn't test tsunami of heights that had occurred in 1896 and 1933. It was beyond their imagination that there could ever be a tsunami as high as those that occurred in 1896 and 1933.

When the whole Fukushima thing started, I found myself running down a sequential checklist: Japan, check. Offshore earthquake, check. Tsunami, check. Nuclear radiation, check.

At this point I knew what had to be next. We've all seen the movies a hundred times.

So how did they manage to avoid waking up Godzilla?

2013-07-25 Reply Admin

da Doctah:
When the whole Fukushima thing started, I found myself running down a sequential checklist: Japan, check. Offshore earthquake, check. Tsunami, check. Nuclear radiation, check.
At this point I knew what had to be next. We've all seen the movies a hundred times.

So how did they manage to avoid waking up Godzilla?

They didn't. Godzilla awoke, became prime minister again, and set about destroying the yen.

2013-07-25 Reply Admin

I think Shoreline refers to the test that occured in March 2011.

2013-07-25 Reply Admin

Mark:
In Mother Russia, you can draw circle ... as long as circle have 4 corners.

In Soviet Russia, circle draws you!

Steve The Cynic · 2013-07-25 Reply Admin

belzebub:
The only thing I can think of is that it must've been REALLY, REALLY old computer, and therefore very SLOW one. Not only slow, but:

Old computers didn't have any graphics acceleration, and often no graphics libraries. There could've been some form of "ClearScreen", "DrawPixel" if you were lucky.

There usually was no floating point processor.

Not many people working with OLD computers were familiar with "new cool" algorithms, like the Bresenham's circle algorithm.

You know that Bresenham's work was done in the late 60s, don't you? So in fact the "new cool" algorithms were first implemented on what is now a catastrophically OLD computer.

belzebub:
So, when a programmer in Czech Republic was working on mainframe in 80's (they was not manufactured after the fall of communism), and he had to write circle drawing function, he would probably do something like this:

Even in the 80s, knowledge of Bresenham's work would have been sufficiently widespread for circles to be blindingly fast and to require no floating point operations. And the Soviet-bloc mainframes were reverse-engineered clones of the IBM System/360 and System/370 mainframes, which date to the late 60s as well.

So in your weird world, these "new cool" algorithms are now nearly fifty years old. And even at the time the supposed events of the story supposedly took place, these algorithms were twenty years old. "New cool" evidently means "twenty years old".

And another thing. I'm disappointed that nobody, not even a self-proclaimed Czech, commented on one thing. Ulrich would not have called a guy in the Czech Republic in the 1980s, because in the 1980s no such country existed. Miklos and Blazej would have been Czechoslovakian, and the mainframe would have been bought from Czechoslovakia.

The standard of TDWTF writing is TRWTF.

2013-07-25 Reply Admin

At 2013-07-24 11:58 :-

belzebub:
Also, as I'm Czech, this is super offensive :)

At 2013-07-24 12:40 :-

chubertdev:
So I'm assuming that we don't have Eastern Europeans here who take offense to stereotypes, even ones as badly written as these.

You see what I did there.

2013-07-25 Reply Admin

Steve The Cynic:
Miklos and Blazej would have been Czechoslovakian, and the mainframe would have been bought from Czechoslovakia.

Even in those days Czechs referred to themselves as Czechs instead of Czechoslovakians. I'm pretty sure Czechs included Slovaks in those days.

Someone claimed to have witnessed a Czech and a Slovak get caught and eaten by a couple of bears. When he said that the female bear ate the Slovak, everyone knew he was lying.

2013-07-25 Reply Admin

How are empty comments even possible?

no laughing matter · 2013-07-25 Reply Admin

Steve The Cynic:
You know that Bresenham's work was done in the late 60s, don't you? So in fact the "new cool" algorithms were first implemented on what is now a catastrophically OLD computer. ... So in your weird world, these "new cool" algorithms are now nearly fifty years old. And even at the time the supposed events of the story supposedly took place, these algorithms were twenty years old. "New cool" evidently means "twenty years old".
And another thing. I'm disappointed that nobody, not even a self-proclaimed Czech, commented on one thing. Ulrich would not have called a guy in the Czech Republic in the 1980s, because in the 1980s no such country existed. Miklos and Blazej would have been Czechoslovakian, and the mainframe would have been bought from Czechoslovakia.

The standard of TDWTF writing is TRWTF.

Lazy reading today, eh?

The article states that this story happened in the late nineties and that they were preparing for Y2K.

So if we believe in this story for a minute, we find out:

Ulrich didn't know of the Bresenham algorithm, as he was only used to computers that were fast enough for a trigonometric solution. A WTF on its own.
The story is still implausible, because there is no reason that a czech company still provides support for soviet-era mainframes.

In fact the wikipedia article i linked earlier states that IBM took over support for those /360 and /370 clones.

no laughing matter · 2013-07-25 Reply Admin

Keith Thompson:
If he was working at the nuclear power plant in the picture, he should already know Czech; that's the Dukovany Nuclear Power Station.

Use the source, Keith!

In the HTML comments: "The fact that it's a Czech reactor is a complete coincidence."

no laughing matter · 2013-07-25 Reply Admin

Taemyr:
Tracy Kidder:
no laughing matter:
Shoreline:
I'm pretty sure this is TRWTF:
They didn’t have a “test control room”...
This is a nuclear power plant, and they can't test what happens in an equivalent environment. The plant will blow up. We are all fucked.
The cautious guys at Fukushima had not only one, but three reserve plants to test what happens when you operate a nuclear reactor in a region prone to earthquakes and tsunamis. The test results showed that the plant will blow up, which could be reliably reproduced in subsequent tests!
No they did not test real tsunami. They might have tested earthquakes of magnitudes that had occured in the past; I don't know if they did. But they didn't test tsunami of heights that had occurred in 1896 and 1933. It was beyond their imagination that there could ever be a tsunami as high as those that occurred in 1896 and 1933.
I think Shoreline refers to the test that occured in March 2011.

Well, no, Shoreline refers to the reactor in the article (which hasn't blown up so far). But indeed i was referring to the events in March 2011.

And a whooosh to Tracy Kidder.

eViLegion · 2013-07-25 Reply Admin

no laughing matter:
* Ulrich didn't know of the Bresenham algorithm, as he was only used to computers that were fast enough for a trigonometric solution. A WTF on its own.

Not knowing an algorithm isn't a WTF. Not knowing an obsolete algorithm even less so.

no laughing matter:
* The story is still implausible, because there is no reason that a czech company still provides support for soviet-era mainframes.

Why do you say there is no reason? If the mainframes are still in use, then there is a market for their support. A tech company from the former soviet bloc would be exactly the kind of company to support it.

Incidentally, what makes you think this mainframe was a 360 or 370 clone? The make/model of machine isn't mentioned.

2013-07-25 Reply Admin

Norman Diamond:
¯\(°_o)/¯ I DUNNO LOL:
And I remember working on a system in the mid 80s that had all the requirements for hard real-time operation, notably the fact that if high-priority tasks are running, low priority tasks don't get a time slice, *ever*.
I used VAX/VMS back then too, when I was in college.
VMS did not work that way. When a low priority task had a page fault, its priority got bumped 6 levels.

Really?!?!? A page fault bumps the pri of a process? Funny I never saw that in the source code, or talked about

Now did something like waiting for a LEF, which would happen on an IO give a 5-6 level boost, sure. But that was only good for a few quanta then it was back down to the base pri. The idea was, a user just hit enter on a command, the IO finished, give him a small short boost to at least get his command started.

If a page fault resulted in an IO to pull something in, then yes the IO completion gave you a boost, but not a page fault. Think about it, if you are CPU bound, and something at a higher pri is CPU bound, you are not getting any time. OK in a rare case you can read below.

And on top of that you can not have a page fault if you are not a running process.

Read this http://www.parsec.com/public/cpuperformance.pdf

2013-07-25 Reply Admin

Steve The Cynic:
You know that Bresenham's work was done in the late 60s, don't you? So in fact the "new cool" algorithms were first implemented on what is now a catastrophically OLD computer.

Oh. So you didn't notice the quotes, right? Right? That means I was being ironic. So once more, for the slow ones:

Even though Bresenham's circle algorithm has existed for some time, MOST programmers (especially czech ones) didn't know it. And they didn't know it because it was perhaps published in some scientific paper (programmers hardly ever read scientific papers) and (this is the good part) there was no easy way for czech programmer to know about it. There was no World Wide Web. Either you read some 10-20 years czech (read as "outdated" by 10 years the day it was printed) book about "kybernetics", or you learned by trial and error.

I even own one czech book about programming (from 70's) which actually contains the "advice" very similar to what I wrote earlier (using sin and cos and very fine step).

Steve The Cynic:
Even in the 80s, knowledge of Bresenham's work would have been sufficiently widespread for circles to be blindingly fast and to require no floating point operations.

As I shown before, no, it wouldn't. Now, it could be true, but in 80's, in soviet block country? Most czech people didn't know english (you generally learnt russian at school), and most czech programmers didn't know english even in 90's.

Now note that I'm still saying "80's" and not "90's" - that's because there was NOTHING you could call "a mainframe" manufactured in Czech Republic (or former Czechoslovakia or short-lived Czechoslovakian Federal Republic) afther the revolution in 1989. So this mainframe must have been from 1980's.

Steve The Cynic:
So in your weird world, these "new cool" algorithms are now nearly fifty years old. And even at the time the supposed events of the story supposedly took place, these algorithms were twenty years old. "New cool" evidently means "twenty years old".

Again, note the "ironic" quotes. And of course, you fail to recognize the main ingredience in a good WTF - basic human stupidity. Just because it was published, or it's known, doesn't mean everyone will use it or use it right. Or, as I've said, it was even pretty hard to really know about it. There was no WWW in th 80's! And for a czech guy, you were pretty much isolated, and your biggest challenge was to even get close to some computer. And if you did, you were stuck with pretty basic sw tools. And by basic, I mean the best you could hope for was some assembler and line-based text editor.

Steve The Cynic:
And another thing. I'm disappointed that nobody, not even a self-proclaimed Czech, commented on one thing. Ulrich would not have called a guy in the Czech Republic in the 1980s, because in the 1980s no such country existed. Miklos and Blazej would have been Czechoslovakian, and the mainframe would have been bought from Czechoslovakia.

You're so smaart :) Only you probably don't know that "Czechoslovakia" means "Czech Republic" and "Slovak Republic" federation. Same as USA, but only 2 states instead of 51. And when the article said "Czech Republic" I just followed, because there WAS Czech Republic in 80's, it was only part of the CzechoSlovakian Socialistic Republic (CSSR). The same way Texas exists in USA.

no laughing matter · 2013-07-25 Reply Admin

eViLegion:
no laughing matter:
* Ulrich didn't know of the Bresenham algorithm, as he was only used to computers that were fast enough for a trigonometric solution. A WTF on its own.

Not knowing an algorithm isn't a WTF. Not knowing an obsolete algorithm even less so.

It's not really obsolete, being still much faster and producing adequate results. Only if you really don't have to care about performance it is acceptable to not know this algorithm.

However Ulrich is programming for very obsolete hardware, so he should have good knowledge about how to avoid wasting resources.

eViLegion:

no laughing matter:
* The story is still implausible, because there is no reason that a czech company still provides support for soviet-era mainframes.

Why do you say there is no reason? If the mainframes are still in use, then there is a market for their support. A tech company from the former soviet bloc would be exactly the kind of company to support it.

Not much of a market, if you think about it: After the end of the soviet bloc corporations (if still in business) then had access to much cheaper and more powerful hardware, so why keep the old ones in operation?

And newly founded businesses had no reason to use those old machines.

eViLegion:
Incidentally, what makes you think this mainframe was a 360 or 370 clone? The make/model of machine isn't mentioned.

I linked to two relevant wikipedia articles in this comment.

Most of the mainframes built in the soviet bloc where built in russia itself or in belarus.

The ES EVM is the only one with a mention of production in Czechoslovakia.

From the ES EVM wikipedia page:

There were even anecdotal rumors among the Soviet programmers, that this supposedly Soviet operating system contained some secret command, which outputs the American national anthem. Today some of the Russian institutes that worked on ES EVM are cooperating with IBM to continue legacy support for both genuine IBM mainframes and the ES EVM systems.

The Circle of Fail

Leave a comment on “The Circle of Fail”