# Differential Operator

1. Dec 13, 2005

### Treadstone 71

This may seem like an odd question, but why are there two different ops for the normal and partial derivatives? i.e., $$\frac{d}{dx}$$ and $$\frac{\partial}{\partial x}$$? I don't see a difference if only one is used, since we are always differentiating wrt a single variable anyway.

2. Dec 13, 2005

### incognitO

think of it....

in classical mechanics, you can have functions like the action $L$ wich is defined in the form

$$L=L(x,y,z,t)$$,

but $x=x(t)$, $y=y(t)$, $z=z(t)$, so

$$\frac{d L}{d t}=\frac {\partial L}{\partial x} \dot{x}(t)+\frac {\partial L}{\partial y} \dot{y}(t)+\frac {\partial L}{\partial z} \dot{z}(t)$$

wich clearly is different from $\partial L/\partial t$.

EDIT:

Sorry, my mistake... the derivative is missing one term. It should be read

$$\frac{d L}{d t}=\frac {\partial L}{\partial x} \dot{x}(t)+\frac {\partial L}{\partial y} \dot{y}(t)+\frac {\partial L}{\partial z} \dot{z}(t)+\frac{\partial L}{\partial t}$$

Last edited: Dec 13, 2005
3. Dec 13, 2005

### Treadstone 71

Interesting. I never encountered those. Then again, I'm not in physics.

4. Dec 13, 2005

### incognitO

The above are functions present in Hamiltonian systems, wich are a big subject of study for mathematitians too... Specially in P.D.E.

EDIT:

Not to mention Calculus of Variations.

5. Dec 13, 2005

### matt grime

one came first, d/dx, and the other is a generalization of it, but asking what d/dx of some object is is strictly different from asking what partial d by dx of it is since the former assumes that the other variables (if there are any) are a function of x too. That is to say that if f(x)=x+y then

$$\frac{\partial f}{\partial x}$$

makes sense but

$$\frac{df}{dx}$$

doesn't

6. Dec 13, 2005

### Treadstone 71

Wait, if $$f(x)=x+y$$, wouldn't $$\frac{df}{dx}$$ make sense since y is a constant? That is, $$\frac{df}{dx}$$ does not make sense if it was $$f(x,y)=x+y$$?

7. Dec 13, 2005

### matt grime

And what if y weren't a constant? come on, put the pieces together, you should be able to correct the obvious mistakes that people make! Dear God.

8. Dec 13, 2005

### Treadstone 71

Fascinating.

9. Dec 14, 2005

### matt grime

Here's another reason for the disticntion.

I give you y, jsut y, now differentiate it with respect to x. What's the answer? dy/dx or 0?

I suppose it is unfair of me to expect you to recognize silly errors from catastrpohically bad ones, not to mention hypocritical perhaps.

Last edited: Dec 14, 2005
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook