Plotting functions anf computing gradient

i am currently trying to plot a function and the compute the graident of f using the jacobian function

Plot the function f (x, y) = x sin(xy) − y sin(5y).

Compute the gradient of f in using the jacobian function

my current code is shown below yet im getting erors

f = @(x,y) x*sin(x,y) - y*sin(5*y);
syms x y
jacobian([x*sin(x,y - y*sin(5*y))])

Error using sym/sin
Too many input arguments.

gradient(x*sin(x,y - y*sin(5*y)))

Answer 1

[Walter Roberson]

sin(xy) is sin(x*y) not sin(x,y)

f = @(x,y) x*sin(x*y) - y*sin(5*y);
syms x y
jacobian(f(x,y)).'

ans = 

[Paul]

Is there any advantage/disadvantage to using an anonymous function vs a symfun?

syms x y
f(x,y) = x*sin(x*y) - y*sin(5*y);
jacobian(f(x,y),[x y])

ans = 

Seems to me a symfun (or just a plain sym expression) is more natural ...

[Walter Roberson]

In this particular case, I was going by the principal of least change to the original code. An anonymous function can be used with jacobian, and the user used an anonymous function, so I showed the small change to the code the user had created.

Sometimes I adjust user code to be better form, but sometimes I just show the minimal change.

More generally, the anonymous function can be used with purely numeric scalar inputs, producing double output, but a symfun would produce a symbolic output, which might or might not be acceptable to the caller.

The posted anonymous function cannot be used with non-scalar inputs. A symfun would work for this computation with nonscalars