solving a physic problem using sage

edit

I think the problem is that you are really asking a physics question, not a Sage question. So anyone trying to answer will first have to make sure your physics is right. Certainly Sage does include several ways to solve systems of differential equations, numerically or exactly - see these Maxima solvers or this well-understood library of numerical methods.

Your first three equations for the balance of the components of force look fine. Your fourth equation, while accurate, is less useful than writing the tensions $S_x$ and $S_y$ as projections of the total tension $T$:

$ \qquad S_x = \frac{x}{\sqrt{x^2+j^2}}T \qquad S_y = \frac{j}{\sqrt{x^2+j^2}}T $

Your fifth equation is incorrect, since it says that the amount of rope on each side of the nail is always equal, which is not true for a sliding inelastic rope. If the rope is of length $L$, then the constraint you want is

$ \qquad y + \sqrt{x^2+j^2} = L $

If you take a second derivative of this constraint with respect to time and substitute it in the first equation, you'll have an expression for the total tension $T$ as a function of $x$ and $j$ and their derivatives. You can use that to eliminate $T$ from the second and third equations. Then you should have three coupled nonlinear differential equations for $x$, $j$ and $y$ without $T$.

At that point you'll have to numerically integrate the equations as a system using for example this method. You'll need to solve for all second derivatives and then introduce intermediate variables for the first derivatives of $x$, $j$ and $y$ in order to write the system in terms of first-order equations.

The problem you're trying to solve is mathematically identical to the swinging Atwood machine, but with one point of support instead of two. While it can be done in Cartesian coordinates, it will look a lot simpler in polar coordinates.

Cheers!