Problem
The equation $x+z+(y+z)^2=6$ defines z implicitly as a function of x and y ,say z=f(x,y) . Compute the partial derivative $\frac{\partial f}{\partial x}$
Attempt
Let $g$ be a auxiliary function defined as follows
$g(x,y)=F(x,y,f(x,y))=0$
Then,
$\frac{\partial g}{\partial x}= \frac{\partial F}{\partial x}+\frac{\partial F}{\partial z}\frac{\partial f}{\partial x}$
and $\frac{\partial g}{\partial y}= \frac{\partial F}{\partial y}+\frac{\partial F}{\partial z}\frac{\partial f}{\partial y}$.
After this my approach is to equate both $\frac{\partial g}{\partial x}$ and $\frac{\partial f}{\partial y}$ to zero but i don't have any logical reason for doing it.
My previous related to this doubt of mine. 1
Please help
Your approach is correct: indeed if you have $F(x,y,f(x,y))=0$ deriving both sides gives you $\partial_xF=\partial_x0=0$ and the same for $y$, and this is allowed because you are "doing the same on both sides".
Now the complete approach: Just substitute $z=z(x,y)$ in the equation and bring everything to the left:
$$F(x,y,z(x,y))=x+z(x,y)+(y+z(x,y))^2-6=0$$
Now derive for $x$: this is a linear operation and thus we can derive both sides of the equation. Moreover $y$ is independent so $\partial_x y=0$ and thus $$\partial_x(F(x,y,z(x,y)))=\partial_xF+\partial_zF\cdot\partial_xz=1+2(y+z(x,y))\cdot \partial_xz=0$$ Thus $$\partial_x z=\frac1{1-2(y+z(x,y))}=\frac1{1-2(y+z)}.$$ Now you either solve this PDE or you substitute $z$ via solving the previous equation (which is just a quadratic equation in $z$).