How do I manipulate infinitesimals? How do I know what assumption is valid?

**ssadi** · May 21st 2010, 06:14 PM

Hello,
I know certain things about small finite increments, like how the general formula for dy/dx is derived, but I have no idea how to manipulate the infinitesimals. The assumptions made seems arbitrary to me, how do I know that it is ok to assume one thing, but wrong to assume in case of others? I mean I can understand that this follows from that assumption and etc, but what made the assumptions valid I have no idea. Can you please explain to me the concept of delta x, or give me links to some resources so that I can understand at least well enough to get sums like these done from scratch:

**CaptainBlack** · May 21st 2010, 10:29 PM

Originally Posted by ssadi

Hello,
I know certain things about infinitesimals, like how the general formula for dy/dx is derived, but I have no idea how to manipulate the infinitesimals. The assumptions made seems arbitrary to me, how do I know that it is ok to assume one thing, but wrong to assume in case of others? I mean I can understand that this follows from that assumption and etc, but what made the assumptions valid I have no idea. Can you please explain to me the concept of delta x, or give me links to some resources so that I can understand at least well enough to get sums like these done from scratch:

There are no infinitesimals in your attachments there are small finite increments (that is what $\delta x$ denotes) which are subject to all the usual rules of arithmetic and common algebra and there are limiting processes converting sums with finite increments to integrals as increments become arbitrary small.

CB

**ssadi** · May 21st 2010, 11:22 PM

Originally Posted by CaptainBlack

There are no infinitesimals in your attachments there are small finite increments (that is what $\delta x$ denotes) which are subject to all the usual rules of arithmetic and common algebra and there are limiting processes converting sums with finite increments to integrals as increments become arbitrary small.

CB

Sorry, i didn't know what delta x was called

Ok, here is my problem, in the following diagram the height (length) of the cylinder:
My calculations: $Rcos(\theta)-Rcos(\theta+\delta \theta)$ = $Rcos(\theta)-R(cos(\theta)cos(\delta \theta)+sin(\theta)sin(\delta \theta)$
$=R(cos\theta-cos\theta+sin\theta*\delta\theta)$
$=Rsin\theta*\delta\theta$ (assumptions $cos\delta\theta$ becomes 1, $sin\delta\theta$ becomes $\delta \theta$
But the book says that the height is R\delta \theta, what am I doing wrong here?

**CaptainBlack** · May 22nd 2010, 01:43 AM

Originally Posted by ssadi

Sorry, i didn't know what delta x was called

Ok, here is my problem, in the following diagram the height (length) of the cylinder:
My calculations: $Rcos(\theta)-Rcos(\theta+\delta \theta)$ = $Rcos(\theta)-R(cos(\theta)cos(\delta \theta)+sin(\theta)sin(\delta \theta)$
$=R(cos\theta-cos\theta+sin\theta*\delta\theta)$
$=Rsin\theta*\delta\theta$ (assumptions $cos\delta\theta$ becomes 1, $sin\delta\theta$ becomes $\delta \theta$
But the book says that the height is R\delta \theta, what am I doing wrong here?

You are computing different things, you are computing the x extent of the ring while the book is computing the slant height of the ring.

CB

**ssadi** · May 22nd 2010, 08:59 AM

Originally Posted by CaptainBlack

You are computing different things, you are computing the x extent of the ring while the book is computing the slant height of the ring.

CB

But a cylinder has no slant height, it is a property of cones, and the book instructs me to use elemental cylinders. How do I know that I have to calculate slant height when it tells me to use cylinders, I calculated what I knew to be height of the cylinders!

**CaptainBlack** · May 22nd 2010, 10:41 AM

Originally Posted by ssadi

But a cylinder has no slant height, it is a property of cones, and the book instructs me to use elemental cylinders. How do I know that I have to calculate slant height when it tells me to use cylinders, I calculated what I knew to be height of the cylinders!

You are going to be computing the mass of a slice of a spherical shell, this is approximately a part of a cone and does have a slant height (and that is what you need to calculate its volume and hence mass.

CB

**ssadi** · May 22nd 2010, 11:14 PM

Originally Posted by CaptainBlack

You are going to be computing the mass of a slice of a spherical shell, this is approximately a part of a cone and does have a slant height (and that is what you need to calculate its volume and hence mass.

CB

If I take this as part of a cone, why do I compute its mass using the formula of a cylinder's volume? And if using the formula for cylinder's volume is ok, why is it not ok to use cylinder's height (which I calculated) instead of slant height?

**CaptainBlack** · May 23rd 2010, 12:37 AM

Originally Posted by ssadi

If I take this as part of a cone, why do I compute its mass using the formula of a cylinder's volume? And if using the formula for cylinder's volume is ok, why is it not ok to use cylinder's height (which I calculated) instead of slant height?

You have not posted the image file showing what you say, so my guess is that you have misconstrued the use of the area formula for a parallelogram for that of a rectangle.

CB

**HallsofIvy** · May 23rd 2010, 01:12 AM

Originally Posted by ssadi

If I take this as part of a cone, why do I compute its mass using the formula of a cylinder's volume?

Because, for constant density, weight is equal to "density times volume". And if density is a variable you choose your "pieces" so that the density is at least approximately constant in each piece- sum the pieces and take the limit as their size goes to 0 giving you of the form "density times dVolume".

And if using the formula for cylinder's volume is ok, why is it not ok to use cylinder's height (which I calculated) instead of slant height?

If you are talking about problem 19, a hemi-spherical shell, each band has bases of two different sizes- part of a cone, not a cylinder.

**ssadi** · May 23rd 2010, 05:26 AM

Originally Posted by CaptainBlack

You have not posted the image file showing what you say, so my guess is that you have misconstrued the use of the area formula for a parallelogram for that of a rectangle.

CB

The diagram is at post #3.

**ssadi** · May 23rd 2010, 05:31 AM

Originally Posted by HallsofIvy

Because, for constant density, weight is equal to "density times volume". And if density is a variable you choose your "pieces" so that the density is at least approximately constant in each piece- sum the pieces and take the limit as their size goes to 0 giving you of the form "density times dVolume".

If you are talking about problem 19, a hemi-spherical shell, each band has bases of two different sizes- part of a cone, not a cylinder.

How do I learn to make a valid assumption for similar sums? Can you give me some website links etc?

**CaptainBlack** · May 23rd 2010, 07:21 PM

Originally Posted by ssadi

The diagram is at post #3.

The term "cylindrical" is being used in a vague and misleading manner. and that diagram is poorly labelled.

CB

**ssadi** · June 1st 2010, 01:52 PM

It was the question, not my own drawing

**CaptainBlack** · June 1st 2010, 03:52 PM

Originally Posted by ssadi

It was the question, not my own drawing

I'm afraid that experience suggests that poor or sloppy wording is not uncommon in texts.

CB

Thread: How do I manipulate infinitesimals? How do I know what assumption is valid?

LinkBack

Thread Tools

Display

How do I manipulate small finite increments? How do I know what assumption is valid?

Similar Math Help Forum Discussions

is this a 'safe' assumption?

The Cantor set and infinitesimals

Infinitesimals and Completeness Axiom

Manipulate big equation

Manipulate Math with Java

Search Tags