3

If we start with:

$$\dfrac{1-\cos(2\theta)}{\sin^2(2\theta)}\;,$$

we can simplify that to , $$\dfrac{1-(1-2\sin^2(\theta)}{4\sin^2(\theta)\cos^2(\theta)}\;,$$

which simplifies further to

$$\dfrac{\sin^2(\theta)}{2\sin^2(\theta)\cos^2(\theta)}\;.$$

Now, I can divide both sides by $\sin^2(\theta)$, (given that it doesnt equal $0$), and arrive at the answer $\dfrac{1}{2}$$\sec^2(\theta)$.
Now, surely the domain resitriction I must apply is all the values of $\theta$ for which $\sin^2(\theta)$ equals $0$, such as $\pi$.
Now this is the bit that makes little sense, if I put $\pi$ into the first equation, I get an undefined answer, but in $\dfrac{1}{2}$$\sec^2(\theta)$, I get $0.5$, which is what desmos plots.

So my question is, what is happening here?

Angelo
  • 12,328
Nav Bhatthal
  • 1,057
  • You might find this question and my answer there helpful for a discussion of removable discontinuities (which is what your question is really about): https://math.stackexchange.com/questions/1525054/why-are-removable-discontinuities-even-discontinuities-at-all/1525156#1525156 – mweiss May 29 '23 at 16:17

3 Answers3

2

What you have calculated, put in more precise words, is this:

For any real number $\theta$ such that $\sin^2(2\theta)\neq 0$, we have $$ \frac{1-\cos(2\theta)}{\sin^2(2\theta)}=\frac{1}{2}\sec^2(\theta). \tag{1} $$

The fact that $(1)$ does not hold when $\theta=\pi$ does not falsify the above result you have shown, because $\sin^2(2\theta)=0$ in that case.

durianice
  • 2,624
  • Ah. So even if it holds when $\theta$ = $\pi$ it doesnt mean anything as it was already "said" that we arent looking at those cases. (Or rather it only holds if we take the limit) – Nav Bhatthal May 29 '23 at 14:10
  • @NavBhatthal Yes, that's correct. Also, the question about desmos is addressed by another answer. – durianice May 29 '23 at 14:11
  • I guess this is the issue with tech like desmos. We can say for sure the two expressions are equal when $\theta$ doesnt equal $\pi$, but thats not to say the limit exists. Thanks for your answer, ill mark it as the correct one. – Nav Bhatthal May 29 '23 at 14:13
  • @NavBhatthal I think you meant the opposite: We can say for sure the two expressions are equal when $\theta\neq\pi$, but the graph also suggests the limit of LHS when $\theta\to\pi$ is also equal to $\frac{1}{2}\sec^2(\pi)$. – durianice May 29 '23 at 14:16
  • Yeah sorry for my lack of good english skills even though I am a native speaker – Nav Bhatthal May 29 '23 at 14:17
2

I want to address what Desmos is doing, as the other answers have explained what's going on algebraically.

Desmos is actually pretty smart when it comes to functions with holes in their domain: it can visually distinguish between "points" and "holes" on the graph of a function. To see this, try clicking and holding on a point on the graph of your function. You should see something like this:

enter image description here

Now drag the point to the left, and see what happens when you reach $x = 0$:

enter image description here

Do you see how the point style changes from a solid circle to an open one, and the point label changes to indicate that the function is undefined there?

Now let's see what happens when you click and drag near $x = \pi$:

enter image description here

Why isn't Desmos displaying this "hole" as an open circle? It's because we can't drag the point exactly onto the point where the $x$-coordinate is $\pi$. Desmos is describing the point using a decimal approximation, and the best we can do is get close to $\pi$.

However, you can confirm that Desmos does actually know that the function is undefined there by explicitly asking (in the Algebra pane) for an evaluation of $f(\pi)$:

enter image description here

Once you know that there are supposed to be holes in the graph of the function whenever $x$ is a multiple of $\pi$, you can manually add them to the graph by plotting a sequence of points and setting the point style to be an open circle:

enter image description here

amends the graph as follows (only a portion is shown below):

enter image description here

mweiss
  • 23,647
1

This is a conceptual question.

Generally, suppose $f$ and $g$ are functions from some set $D$ to the real line $\mathbb{R}$. Let $E$ be the subset of $D$:$E=\{x\in D |g(x)=0\}$. When you try to simplify $\frac{f(x)}{g(x)}$, you need to calculate it on the set $D\setminus E$ (where $g(x) \neq 0$).

Here's a simple example, let $f(x) = x^3$, $g(x) = x^2$, and $x \in \mathbb{R}$. Then, $\frac{f(x)}{g(x)} = x$ (for $x \neq 0$), the right hand side $x$ is defined on $\mathbb{R}$, but $\frac{f(x)}{g(x)}$ is defined on $\mathbb{R} \setminus \{0\}$. There is no essential difference between this example and the question you asked.

PPP
  • 346