Hacker News new | past | comments | ask | show | jobs | submit login
X-3/x-1=x-4/x-5 (bing.com)
38 points by anigbrowl on July 16, 2009 | hide | past | web | favorite | 31 comments

In other words, bing now has some calculation abilities. MS is running an ad with this particular calculation suggested as a Bing search and a background photo of a blackboard.

Well, as you see, it works. And (I presume) they would like you think about the fact that throwing the same equation at Google does nothing useful (inexplicably, the first result was a page about BSD).

But I wonder how many users of this feature will be familiar with the rules of operator precedence? I admit I read it as:


--- = ... << I can't format it right but you get the idea


After testing it with Google, I went to Alpha, which (unsurprisingly) handled it with aplomb, giving it to me with proper notation as well and making me aware of my mistake. So I tried it as (x-3) / (x-1) = (x-4) / (x-5), which Alpha also handled with Aplomb (11/3 if you are lazy).

However, inputing the latter into Bing (without or without spaces for padding) gave no calculation or result, but just a bunch of (mostly unrelated) search results. Seems rather counter-productive on MS's part.

Apologies if this is excessively trivial.

Even weirder...

(x-3)/x-1=(x-4)/x-5 works OK (ie Bing treats it as math); indeed for (x-3)/x-1=x-4/x-5 Bing properly gives both solutions 1/2(5+/-(29^0.5)).

But it absolutely Does Not Like parentheses in denominator (eg x-3/(x-1)=x-4/x-5) on either side of the equation. Clearly, I need to get out more.

google of course now links to this page first!

For some reason, it only works for certain combinations of countries and languages. E.g., it works for Argentina: http://www.bing.com/search?q=x-3%2fx-1%3dx-4%2fx-5&setmk... but not Chile: http://www.bing.com/search?q=x-3%2fx-1%3dx-4%2fx-5&setmk...

And not for Turkey..

Please don't post homework problems to this newsgroup.

Oh dear, I hope that's a joke :-/ Maybe I should have done it as a 'tell HN' type post. I thought people would be interested in the depth of the parsing tree.

Yes. I thought "newsgroup" would be a giveaway. :-/ No usenetters here, eh?

'course, people may just not like the joke.

I had a comp.lang.lisp flashback, so I'll say you did a good job :D

FWIW: http://www.numberempire.com/equationsolver.php says -1/4. It will only give 11/3 if you explicitly put parentheses: (x-3)/(x-1)=(x-4)/(x-5) . So I believe Bing's result is correct. Based on elementary school order of operations, the implied parentheses on the original equation are: x - (3/x) - 1 = x - (4/x) - 5.

I don't see the result in Bing.

alpha direct link: http://www.wolframalpha.com/input/?i=(x-3)/(x-1)%3D(x-4)/(x-...

That's actually a different question than http://www.wolframalpha.com/input/?i=x-3%2F+x-1%3D+x-4%2F+x-...

Bing's formatting of the results doesn't quite draw your eye to it - look for "Calculation" immediately below "All Results"

Are you outside the USA? I'm in Taiwan and Bing isn't calculating it for me either. It's just returning a bunch of normal search results in Chinese. Maybe calculation only works for queries coming from inside the USA.

Well, I'm in India and it does display the calculations for me.

I'm in Croatia and it's not returning calculations for me. It seems only select locations got this enabled in the first run...

That's weird, I didn't see it either this afternoon, but now I do.

Wolfram Alpha also isn't sure what to do with this:


It's yet another real world question that's come up in conversation that I've wanted to know the answer to, and which Wolfram Alpha would be really awesome if it could handle, but sadly cannot.

Washington, DC is the only US capital.

Surely state capitals count as capitals in the US?

Those are state capitals, not US capitals. Though I guess you could make the case that America has 3 capitals: U - S - A!

An interesting, apparently unintentional, side effect is the exercise in the precedence rules.

Search engines (bing, google, wolfram) get the precedence but are finicky about calculations. Humans (presumable anyway, from the earlier comments) do the calculations but sometimes overlook the precedences. I think it is rather funny.

Live! search could do these types of calculations for sometime. It could even handle solving quadratics.

Live search had an integrated exclamation point in the name?

x = 11/3

No, x = -1/4 is correct. Perhaps you should look up the order of operations.

He's not all wrong. In algebraic notation division is written as a fraction, and has a lower precedence. So it's a bit ambiguous.

I'm getting x = 11/3 as well.

My workings:

(x-3)/(x-1) = (x-4)/(x-5)

(x-3)(x-5) = (x-4)(x-1)

x^2 -8x + 15 = x^2 -5x +4

-8x + 11 = -5x

11 = 3x

x = 11/3

FWIW, wolfram alpha agrees with me.

You changed the equation.

You started with

(x-3)/(x-1) = (x-4)/(x-5)

which is not the same as

x - 3/x - 1 = x - 4/x - 5

, the equation in the submission.

Haha, thanks - I guess the problem is in my parsing :)

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact