But apart from plain old errors, there are also some bugs and Easter Eggs hidden in all versions of Windows. Care to discover 5 of them?

## 1. Naming & Renaming Folders

For the first one, try to create a folder named CON and you will see the following thing happening:

The same thing happens when you try any of the following names:

PRN, LPT1, LPT2, (…), LPT9, NUL, COM1, (…), COM9, and CLOCK$

All of the above names are reserved device names, which cannot be used as file names or folder names, regardless of the file extension. This is a relic from DOS, which has made its way through all versions of Windows, including Windows 7.

Source: Wikipedia

## 2. Advanced Calculations

Let’s do some basic maths together. Please use your head first. Highlight the next two lines to see the respective answers.

What is the square root of four? It’s two, isn’t it?

And what do you get if you subtract two from two? Zero, right?

Now let’s do the above calculation using the Windows calculator. Type in 4, take the square root and from the result subtract 2. What do you get now?

See why they don’t want you to use calculators at school?

To my knowledge, there is no explanation for this result. The operation returns different numbers in different versions of Windows, so it’s possibly a bug by design or an Easter Egg and not some weird functional error. What was your result?

## 3. Notepad Bug (XP Only)

This bug no longer works in Windows Vista or Windows 7, but if you’re still running Windows XP, give it a try.

Launch Notepad and type the following sentence: *Bush hid the facts*

Now save the file as anything you like, close it and open it again. What do you see?

If you did this in Windows XP, you probably see some weird unicode characters or Chinese characters like in the screenshot above. By the way, I have to admit that I ‘forged’ the screenshot because I no longer run Windows XP.

The explanation for this bug lies in the Windows function ‘IsTextUnicode’. When a text file is encoded in Windows-1252 it is interpreted as UTF-16LE, resulting in the so-called mojibake. It means that Notepad recognizes unicode characters representing Chinese characters and translates them back into Chinese characters.

Source: Wikipedia

## 4. Microsoft Word Trick

Open Microsoft Word and type the following: *=rand (5,10)*

What happens is that Word creates 5 paragraphs, each containing 10 identical sentences. The sentence will vary, depending on the primary language of your operating system. On English Windows, you will see this sentence: *The quick brown fox jumps over the lazy dog.*

This is not a bug, it’s a feature! Many of you certainly know that the above sentence contains all letters of the alphabet. In this case it’s nothing more than a dummy or placeholder. Depending on the numbers you pick, you can make it appear in many more paragraphs and copies. The trick is also known as *=rand (200,99)*.

## 5. Solitaire Bug (XP Only)

Here is another trick that apparently disappeared in Windows 7. Please try it if you have Windows XP.

Open Solitaire and click the following key combination: *[ALT] + [SHIFT] + [2]*

When you click the above key combination, the game ends right there and you see the cards jumping to the front in waves, as they do when a game is completed successfully.

What did you see when you tried to reproduce the above bugs and what are your favorite Windows Easter Eggs?

Image credits: Antonov Roman

In Minesweeper on XP, using Win-D would show the desktop, but also stop the clock on the game. My shortest time for all the levels were one second.

Ha, didn't know about that bug. Thanks, Nick!

The notepad bug also works in Windows 10(using build 14393.576) if you type "Diskmark" (without quotes), copy it like a million times and then save it, if you open it again, you will see that "Diskmark" turned into "????".

Thank you for sharing that one, Michal!

This is what I got when I typed "=rand (200,99)": https://cdn.discordapp.com/attachments/233291944233664512/237644084359987200/unknown.png

Isn't that what you expected, Connor? Or are you concerned about the product activation failed message on top?

The winblows calculator is messed up. Only windows calculator will incorrectly calculate the sqrt of a number that when multiplied to itself results in a who integer number and not a floating point. This does not happen in Unix, Linx, OS2, and even Mac os10. Example 2x2=4, now this is NOT 2.1x2.1 or 1.99x1.99 No, it is exactly 2x2.... As mentioned above this happens with 2x2 or 3x3 or 4x4, etc.... sqrt of 2 whole numbers will result in a whole number and not a floating point. Even entering 2.0000x2.0000 = 4.... Every calculator I own, from a Casio to a Texas Instruments all correctly calculate the value of sqrt numbers correctly, hence sqrt(4)=2, not 1.99999.... and when you enter sqrt(4)-2 you get 0 the correct answer. This is validated. In scientific mode you get 0e0, and in engineering mode you also get 0e0, now Texas Instruments cannot be incorrect in this mathematical calculation can they? After all we tried it in normal, scientific, and engineering mode and all of them gave the correct answer of 0 when entering sqrt(4)-2. I would agree to the floating point if in normal mode the sqrt(4)-2=0, but if enabling floating point calculations by going into scientific mode or engineering mode the answer to sqrt(4)-2=-1.0682.... I would then agree that the sqrt(4) is a floating point answer, but this is not the true answer unfortunately. You can even test this by putting the calculators into scientific mode and engineering mode, (Try it on any TI-89, TI-99, etc)

Now here is the fabulous world of Microsoft, they do not think the same, they think the sqrt(4)=1.999999 and claims the answer is always a floating point number(the number i posted is not exactly what they have as an answer, I am just making an example so please refrain from posting there answer).... Now if you type in there crapulator sqrt(4)-2 you would think that due to the facts, the logical true answer would be 0 correct? They do not see that fact though and claim the answer is -1.068281969439142e-19(in standard mode) and . -8.1648465955514287168521180122928e-39(in scientific mode).... here is where they are flawed even more. I will explain why. If in standard mode the answer was -1.068281969439142e-19, then it will also be that same number in scientific mode. Again, verifiable on all TI calculators by just entering a number that cannot be obtained by multiplying two whole numbers like 2x2 or 3x3... etc. Try sqrt(5) just for an example, you will receive the exact number in normal mode, scientific mode, and engineering mode, one being a longer floating point though, but the answers are the same in either case. Even in Winblows Crapulator it will correctly display that. And if that was a different value in each mode it would have displayed it just the same. But Winblows Crapulator does not report back the same values when in standard or scientific mode when calculating the sqrt(4)-2, it thinks the answer is different in standard mode then it is in scientific mode, and they stick to this answer as being the true answer..... That is just downright crazy.

Now for the ones that believe that Microsuck Winblows Crapulator is correct and not flawed, and that Microsuck is never wrong here... Just do some astronomical mathematics.... that is the proof you will ever need..... after all 0.000001 and 0.00001 or 0.0000001 will result in a telescope RA-DEC position to be way off, like by billions if not trillions of light years. If you think i am joking and just blowing smoke out my rear, just go buy you a Celestron NexStar 130SLT, or any other telescope with computer tracking will do, even manual tracking will work but just harder to precisely track a celestial body. Just use Microsucks Winblows Crapulator to do the calculations for locating a celestial body, find one that will need sqrt numbers that are whole and not floating point to locate and track. There are plenty of them out there, the NGC2000 book has hundreds of pages of them. Try it and come back and tell me if you were able to correctly locate and track anything. Good Luck...

P.S. At the Panter Hall Observatory(Located in Austin Texas at the UT campus) the students and also the staff do not use any of the Microcrap Winblows Crapyalaters, We use things like the Texas Instruments, or Casio, or even some cheep ass dollar general calculator that is more accurate then the Slopware that Microphuck Wankerblows Crapavaders was made with. The answer they produce has never been incorrect, EVER, unless the wrong answer was a result of human error entering data incorrectly.

There... I am done... Proof was shown...

The problem with Windows Calculator was totally caused by programmer's laziness / incompetence. The square root function that they wrote always give approximate value for every numbers, even though numbers such 1, 4, 9, 16, ... should give whole number.

When they display 2 as the square root of 4, it's not really 2, it's a number that's smaller than 2, but almost 2. That's why when we substract 2 from it, it resulted in negative number.

Even Windows 10 calculator still produce this result. They didn't bother to rewrite their square root function. They use legacy code from two decades ago.

Thank you for the explanation, Eterni!

Wow, you ARE a real blow hard, aren't ya?! That post just got worse and worse over time, and by the end you sound like a dyslexic *beep*. Get off your high horse 'cos even though MS could have done better, you really don't have a clue about binary and floating point representation despite all your big talk about TI's calcs, astronomy etc.

Just to provide an explanation for the calculator thing... Computers understand only binary (base 2). And the kind of math that humans find easier is decimal (base 10) (since we were born with 10 fingers and all that). Unfortunately, if a decimal terminates in base 10, it may be a non-terminating decimal in base 2, and hence is approximated. For eg, 1/10 in base 10 is just 0.1, but in base 2, it's something like 0.000110011001100110011... i.e. recurring. If you approximate this off to the first 9 digits, it results in 1/16 + 1/32 + 1/256 + 1/512.... = 0.099609375. This is true for a lot of numbers, which results in the kind of "error" you have talked about.

P.S. Nice article :)

Good explanation, Havan, unlike the incoherent ranting of the fellow above who likes to pretend on the interwebs that he's some sort of astronomer or something.

Hey guys i am running windows ten with word 2010 and i typed it in and it came up with around five paragraphs about the insert tab.What is that for?

The problem with the calculator is a bug that microsoft forget and it's totally normal!!!

you should remain from commenting on computer science if you don't know anything about it!!

In every math calculator, only the software can show up the correct decimal value.

See IEEE754 for to understand how math chip are working!!!

that something as a programmer I do know!! I'm a specialist, obviously, not you!!

Thank you for weighing in, Larry. Glad you're here to educate us.

As you correctly pointed out, I'm not a computer science or math specialist. My specialty is writing and editing. Of course, we all try our best and make mistakes. It's human, you know? :)

By the way, the calculator issue persists in Windows 10. Must be hard to fix!

Your also a douche, obviously!!!

Only a real douche doesn't know the basic difference between "you're" and "your".

P.S. Larry got it absolutely right.

the calculator bug holds good for sqrt of any number minus the answer..

ex:sqrt16 - 4 , sqrt36 - 6 , sqrt81 - 9 all gives random results..

Thanks for the update, Tejas!

great article!

type Q33N in notepad and change the font to size to 72 and font to wingdings

you will see the secret about world trade center attack.

Haha that weird result from the calculator doesn't only happen "virtually". Some real scientific calculators don't even know how to subtract specific numbers...

Neat article!

That's pretty messed up.

As for naming & renaming folder, when I type in "con/prn/lpt1/lpt2/com1/com2 & other reserved device name", what I get is just "New Folder" folder without name. The dialog box does not even appear.

Might be different depending on the Windows version or edition. I took the screenshot in Windows 7 Professional.

If you like these Easter Eggs, Google also has a few Easter Eggs to look at too.

Google Translator - http://www.jackcola.org/blog/128-the-best-google-translator-easter-eggs

Google - http://www.jackcola.org/blog/128-the-best-google-translator-easter-eggs

Good thing I mostly use Linux. I only need Windoze for Netflix.

i was saving a image of chromes about:about by the name of "Chrome about:about" and it showed me the error that the file cannot be saved by this name...i checked with different options contain : but nothing work... i guess it happens as it'd bug cmd while changing directory and all...

For the calculator trick, it seems to do that with any square root that you subtract the answer from. If you take anything like square root of 25 minus 5, square root of 16 minus 4, etc. it does that weird random answer.

The notepad bug can still be faked in win7: save it as ANSI, and then in the Open... screen, choose Unicode.

Wikipedia also has a detailed explanation of the "CON" device bug.

http://en.wikipedia.org/wiki/Concon

On prior Windows version directly accessing a path like

`C:concon`

or`C:connul`

would even crash your system because Windows would ignore the initial path identifier and try to access the device directly.I'm not sure if it works on W7 but on XP if you set your screensaver to marque "Who Cares?" and left it a month then it would change to "I do".

Lol XD

That's awesome!!!

The algorithm that MS uses to calculate the square root of a number will almost always result in somewhat inaccurate results (1x10^-19 for 32-bit systems and 8x10^-39 for 64-bit systems). Since Calculator does not display 20 digits, it rounds the result to 2, but it maintains the calculated result for further use. Thus, when you subtract 2 from the result, you are not subtracting it from 2 but from 1.9999999999999999998931718030560858 (approximately) on a 32-bit system.

This rounding error tends to be introduced when attempting to perform binary calculations on decimal numbers. It happens when the number systems don't have common boundaries (binary & hexadecimal have these common boundaries, but the decimal system does not share common boundaries with either of them).

Thanks a lot for the explanation, Bruce! :) That makes a lot of sense.

but then what about any other calculation with regards to square root of 4? like squrt(4)+1 gives 3 and not any other number. its only squrt(4) -2 which gives wierd numbers. :/

Good point. Maybe they forgot to correct the algorithm mistake in this case. Or it's intentional, say a reminder to never trust a computer 100%? Could be humor.

In this specific case, it's an issue with the square root algorithm indeed. But usually, this kind of error is due to "floating point" representation. We express a number in 1.234*10^45 while the computer will have 3.46383349*2^79. The computer uses powers of 2 rather than powers of 10. This means that rounds numbers in decimal might not be round in binary. That's why the approximation errors can sometimes propagate and become visible.

Here, we have 4 which is a round number and its square root is a round number in binary too, so it should work.

I have seen a case where this kind of problem ended up costing thousands of dollars in the company billing system.

Does anyone know about the 'Clock' animation file that can be found under

My ComputerC:WINDOWS. It's a relic in Windows XP. Not sure if it's still their in Vista and 7.I don't. What is the file called and where is it located in Windows XP? I have Windows 7, curious whether it's still there.

It used to be "clock.wmv", I don't think it's included any more.

I have C:WindowsCLOCK.AVI on my WinXP desktop.

Yeah, I remember it as AVI now too. It's not in the Windows folder in W7 though.

If I recall correctly, the "Bush hid the facts" bug works for any sentence with four words in a four-three-three-five letter scheme.

u indeed recall correctly

I just did it with Word 2011 for Mac and I got the sentences.

Heh, goes to show that the Mac team at MS Office is still at the pre-2007 stage feature-wise.

that's it just working

Whoa, are you still on Word 2003? I know Word 2007 and 2010 does not use that sentence any more.

i did on word 2010 and it gave 5 paragraphs about Insert tab !!

Thanks for sharing what it does in Word 2010!

I got paragraphs about the insert tab and galleries in Word 2007 as well. I think that Word help files are lonely and this is a ruse to try to get someone to read them. I fell asleep after the first 4 words though.

Have to admit that I was on Word 2003 until I updated to Windows 7 recently. Now I don't use Word anymore and I can tell you that this doesn't work in LibreOffice. :)

More eggs {xp} can be found here...

http://home.comcast.net/~Suppo...

More eggs {xp} can be found here...

http://home.comcast.net/~SupportCD/XPSecrets.html

Win7 Home Premium 64-Bit: -8.1648465955514287168521180122928e-39

Win7 Home Premium 64-bit: -1.068281969439142e-19

Probably depends on Locale too, dunno, mine is English Australian

The result changes with the view you have set on the calculator.

32-bit too

Hi. This is not exactly a bug. This type of surprising results will happen if you compute the difference of two numbers which are almost identical:

As a rule, the square root of a number cannot be computed exactly. A program determines an approximate value for any square root. here, the square root of 4 was found to be almost 2, and was displayed as 2, but the stored value was not exactly 2, so, of course the difference with 2 yield the rounding error: a 0.0000....(put here 18 zeroes)...1068....

Rounding errors are worse when substraction is used on large numbers, on a system which represents number as Floats, which means all mainstream computer programs.

Avoiding this necessitates the use of structures called BIGNUMS in Lisp/Scheme...

This is the reason why the equality of two non-integer numbers a and b should never be tested with the test "b-a=0", but with something like "b-a<e" , where e is the maximal admissible rounding error.