# Ask Google this impossible question, get web filth as a reward

An odd bug in Google's search algorithms appears to be benefiting XXX-rated websites. Searches for impossible pages - such as the contradictory search term -4^(1/4) which means "Find me pages containing a 1 next to a 4, but which do not contain a 4 - return web-page results liberally sprinkled with links to online grumble flicks …

#### Either that, or,

the search engine is just thinking "that request is a perversion of logic... perversion. Okay!"

#### Re: Either that, or,

Careful with those contradictory search terms then.

You could get Total Perversion and blow a hole in the space/time continuum...

#### Re: Either that, or,

Surely it's just the equivalent of the (already self aware, let's face it) google servers saying:

"That thing you serached for does not exist, but this is what pretty much everybody else is looking at on the internet at the moment so I presume you will want to see it too...."

#### Re: Either that, or,

Is it something to do with searching for two girls but only one cup?

#### Re: Either that, or,

Er actually -4^(1/4) means the fourth root of -4, i.e root 2: 1.414 etc and that is exactly what you get.

I don't know what you have been smoking.

#### Bit of a disappointment...or relief

that the sentence" 1 next to a 4, but not containing 4" is NOT some obscure slang for some weird group action bang....

#### Re: Bit of a disappointment...or relief

Relief - Hand relief to be exact.

Positive searches use different likenesses so

1 next to 4 could equate to one next to four or fore as in solitary foreplay => hand relief

The not containing part is probably literal so -4 means just that, although it might be difficult to find any web page that doesn't have the digit 4 in it but hey it's only an example.

In short and seriously - positive searches are flexible whereas negative ones are strict thereby leading to results for impossible searches, as to why those results should be primarily porn sites, well, why not. According to knee jerk politicians and brain dead DM readers even the most innocent of searches return links to porn sites so the impossible ones might as well do so too.

#### Re: Bit of a disappointment...or relief

"So you're saying it's no longer obscure?"

SMOOTH !!! LIKE IT!

#### I am bitterly disappointed..

In el reg these days, once upon yore I could rely on V central to provide me up to date information on my kitten vid accessibility before the mainstream press but now I find that you are days, (days!) behind them. How can the SMH (Syd Morning Horrible) publish this story before you good folk??

Standards are slipping..

#### according to my parsing...

the "-4^(1/4)" is mathematical symbology for minus 4 to the power of 0.25 and is a perfectly legal term to have in a search string...

#### Re: according to my parsing...

Windows calc says "invalid input" when I try it. Apparently doesn't know about the 4th root of negative numbers. :-)

But, yes, that was my first thought as well and I often use Google as a primitive calculator / unit convertor when I can't be bothered to dig out a calculator.

#### Re: according to my parsing...

And the answer(s) to -4^(1/4) is +/-sqrt(2)i from my memory, where i is the imaginary number equal to sqrt(-1). There is also 1 +/- i.

#### @Lee

Calc is the wrong tool for the job. For formula's you'd really want to grab Microsoft Mathematics (download link) which can easily build and parse formula's. Then you'll get -1.4142135623731 as the answer (minus the square root of 2).

For a super cheesey movie (though I think its kinda funny) check this Youtube video on what Math 4.0 can do for schools. It'll also give you a good impression as to what Math 4 actually is.

#### Re: according to my parsing...

Only in BASIC.

In *real* programming languages, the "hat" is the bitwise EOR operator; or means "beginning" in a search pattern, or turns a "match any" pattern fragment (such as /p[aeiou]t/ which matches pat, pet, pit, pot or put) into "match anything but" (/[^0-9]/ matches any non-digit).

#### Re: according to my parsing...

Bing gives -(4^(1 / 4)) = -1.41421356, Yahoo doesn't know and Ask Jeeves is "currently experiencing an unusually large number of Web searches. Please try your search again."

#### Re: Ask Jeeves is "currently experiencing an unusually large number of Web searches"

Yeah right! I find it hard to believe they have had a large number of searches this side of 2002

#### Re: according to my parsing...

Yeah, I've fallen foul of trying to use ^ for power in C.

#### Re: according to my parsing...

I get -4.74999..... as follows:

brackets have highest priority; (1 / 4) = 0.25

Then the unary minus gives -4

Then the exclusive or (yes, you can have fractions in binary; 0.25 decimal = 0.01 binary) gives, as near as damn it is to swearing, -4.75 (it's actually a recurring fraction; when you flip the bits and add one, you get 100.10111..... which is close enough to 100.11. In fact if you actually used an infinite number of digits, they would actually be equal; the difference between them must have an infinite number of zeros before the 1, and therefore must be equal to zero).

#### Re: according to my parsing...

There you go. -4^(1/4) simplifies to 1+i, which, in the form "one plus an imaginary root", is actually a pretty fair definition of pornography.

#### -1.4142 or 1+1i,

depending on whether you calculate (-4)^(1/4) or -(4^(1/4)). I don't even know what the rules should be, should the - be considered as (-1)* or as part of the number?

#### Re: according to my parsing...

100.10111..... IS 100.11 (not just close enough) in binary, in the same way 0.9999.... IS 1 in denary (decimal). It's just two different ways of representing the same number in these largely useful but very slightly broken ways of representing numbers.

#### Re: according to my parsing...

Strictly speaking, there are going to be four solutions with one in each quadrant of a plane with real numbers along the abscissa and imaginary along the ordinate. The solutions are (1,i), (-1, i), (-1, -i) and (1,-i).

#### Re: -1.4142 or 1+1i,

Remember your BODMAS from early schooling. Brackets, Powers(Of), Division, Multiplication, Add, Subtract. The Bing result is correct, and interprets it as - (4^(1/4)), which is a +/- real result. The other interpretation gives +/- Root2 +/- Root2 i (there are four roots, with evenly spaced arguments. By the way, I'm a little rusty too. The unary minus is interpreted as (-1) * 4 by Mathematica, so I can see the confusion here.

#### You lot never fail me!

Only El Reg commentards could hijack a perfectly good article about porn by turning it into a debate about parsing mathematical formulae!

Have a pint, and keep up the good work!

#### Re: according to my parsing... @Ralph B

Well done that man! Almost cost me a keyboard, but the cup wasn't quite tilted enough!

#### Re: according to my parsing...

And in fact if you put that into Google, it does tell you -(4^(1 / 4)) = -1.41421356, having apparently decided to apply the minus last. It also returns a site titled "Free real sister sex with brother - All best porn!". So if you like incest with your maths, it's perfect.

#### Fixed?

When I type in "-4^(1/4)" I get the calculator and a result of -1.41421356237 and the "1 2" minus 1 or 2 returns nothing. So it seems they've fixed it already. Either that or it was a non-story in the first place.

#### Re: Fixed?

-4^(1/4) is the 4th root of -4, which is not -1.4142.... The 2nd (square) root of -4 is 2j (where j is the square root of -1)

I'll leave the 4th root to your imagination and a bit more calculator fondling if you're in the mood.

#### Re: Fixed? @frank ly

Nice condescension, but calculator fondling gives the same result on qalc and my Casio. Forgot all my complex numbers stuff years ago so dunno whether that's the right way for them to be reporting it.

#### Re: Fixed? @frank ly

Check what the screen shows after you do "minus 4". Windows calculator, for instance, just shows four.

Now press Enter. You get "proper" -4. Then do the power. You get "invalid input".

#### Re: There are multiple complex roots

But doesn't the ^ have higher prio than the minus in front of the 4? In which case it's -(4^0.25) and no complex number involved.

#### Re: There are multiple complex roots

There are always 4 roots of a quadratic root (^1/4). In this instance the answers are:

1 + i

1 - i

sqrt(2) . i

- sqrt(2) . i

#### Re: There are multiple complex roots

no, because the minus here is a sign indicating a position on the number line not an operator

#### Re: There are multiple complex roots

"no, because the minus here is a sign indicating a position on the number line not an operator"

WHAT! Unary operator is unary. Symmetry operation around point 0.

#### @Annihilator Re: There are multiple complex roots

A quick check will show that +/- sqrt(2).i doesn't give -4 if raised to the 4th power.

The other two roots are (-1+i) and (-1-i). If you draw them on the complex plane, the symmetry is obvious (which is how I figured it out).

#### Re: There are multiple complex roots

(1 + i)^4 = (2i)^2 = -4

(1 - i)^4 = (-2i)^2 = -4

(sqrt(2). . i)^4 = (-2)^2 = 4

(-sqrt(2) . i)^4 = (-2)^2 = 4

#### Re: @Annihilator There are multiple complex roots

@frank ly - bugger! School boy error of assuming root(i) was -i :-(

#### Re: There are multiple complex roots

Hey now, we electrical engineers are just keepin' it real...

#### Q = a_0 + a_1 * i + a_2 * j + a_3 * k

**j** is for the Quaternions, as is **k**

Please use **i** as intended.

#### Re: There are multiple complex roots

"Operator J! What number do you imagine you will be connected to?"

#### Re: Fixed?

Me neither - tried all three and it's either fixed, never existed or perhaps only affects certain configs. Has *anyone* commenting on this thread had actually tried it?

#### Re: Fixed?

No I think it must be fixed, I tried he contradictory terms "bankers" and "nice guys" and I didn't find one p0rn site.

Mines a dity old mac (not the computer)