W|A calculation link

How is this? I'm getting $(-\tan(x))$. Here's my attempt:

Let $u=\sin(x)$ and $v=\cos(x)$. Then, the derivative we seek is,

$$\frac{\mathrm dv}{\mathrm du}$$

Using chain rule, we have,

$$\frac{\mathrm dv}{\mathrm du}=\frac{\mathrm dv}{\mathrm dx}\cdot \frac{\mathrm dx}{\mathrm du}=(-\sin(x))\cdot\frac{1}{\cos(x)}=(-\tan(x))$$

I can't find my flaw. Please help.

Here's the W|A link (shortened by Bit.ly) if anyone wants to verify.

  • 2
    $\begingroup$ dx/du is not quite the same as the reciprocal of du/dx. $\endgroup$ – user217285 Jun 21 '15 at 22:51
  • $\begingroup$ @Nitin, please provide sources to support your claim. Also, can you post an answer as to how one can manually get the answer as $\cot(x)$ ? $\endgroup$ – tom_cruise Jun 21 '15 at 22:53
  • 1
    $\begingroup$ @tom_cruise It's a basic differentiation rule. If I have the function $y=x^2$, then $y'=2x$, and $x'=\frac{1}{2\sqrt{y}}$; the two are not inverses. $\endgroup$ – HDE 226868 Jun 21 '15 at 22:55
  • 7
    $\begingroup$ Actually, HDE, in your example, we see $x'=\frac{1}{2\sqrt y}=\frac{1}{2x}=\frac{1}{y'}$, so it does hold. $\endgroup$ – Alex Mathers Jun 21 '15 at 23:01
  • 2
    $\begingroup$ WA is interpreting the $d$'s as parameters. So the application views $d \cos x/d\sin x$ as $d \times \cos x$ divided by $d \times \sin x$, which is $\cot x$. Your answer is fine. $\endgroup$ – Mark Viola Jun 21 '15 at 23:02

WA interprets the $d$ as a constant, so they cancel out. The resulting fraction, $\frac{\cos x}{\sin x}=\cot x$.

  • $\begingroup$ Damn! Thanks a lot! $\endgroup$ – tom_cruise Jun 21 '15 at 23:00
  • 2
    $\begingroup$ Wish I had made that comment. ;-)) $\endgroup$ – Mark Viola Jun 22 '15 at 0:08

To do the calculation a more plausible way, $$ \frac{d}{d\sin{x}}\cos{x} = \frac{d}{d\sin{x}}\sqrt{1-\sin^2{x}} = \frac{d}{dy}\sqrt{1-y^2}, $$ by writing $y=\sin{x}$ (mere replacement of symbols, doesn't mean anything more) $$ \frac{d}{dy}\sqrt{1-y^2} = -\frac{y}{\sqrt{1-y^2}} = -\frac{\sin{x}}{\cos{x}} = -\tan{x}, $$ so your answer is correct.

My guess is that W|A is misinterpreting your input: for example, this doesn't do what you would expect either. It's probably just interpreting $d$ as a constant, so it thinks it is just doing fraction arithmetic.

  • $\begingroup$ I think you meant $(-\tan x)$ and not $(-\cot x)$. Anyway, thank you! I can't accept both answers so I'm upvoting this one instead. :) $\endgroup$ – tom_cruise Jun 21 '15 at 23:04
  • $\begingroup$ Ah yes, thank you. $\endgroup$ – Chappers Jun 21 '15 at 23:08
  • 1
    $\begingroup$ This is perhaps plausible, but such calculations should always be taken with care. “Derivative of a function WRT another function”, as such, doesn't really make any sense unless the latter is explicitly a function of a function (i.e. a functional). $\cos$ is not really a functional, it's just a real-valued function. $\endgroup$ – leftaroundabout Jun 22 '15 at 14:07
  • $\begingroup$ @leftaroundabout Agreed, although it should hold whenever the chain rule does (that being essentially what I have used, although I have not said so explicitly). $\endgroup$ – Chappers Jun 22 '15 at 15:15

To find $\frac{dx}{du}$ we see $x=\arcsin u$. Then,

$$\frac{dx}{du}=\frac{1}{\sqrt{1-u^2}}=\frac{1}{\sqrt{1-\sin^2 x}}=\frac{1}{\cos x}$$

so that step isn't wrong like Nitin said. I don't see why your solution is wrong, I'd love to see insight anybody else has.

As some others have said, W|A must just be canceling the $d$'s. Your answer is A-okay!!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.