Whenever you see a simultaneous equation you should think of overlapping curves or surfaces. If your equation has two variables (x, y) then it is a curve, if it has three variables it is a surface, if it has n it is an (n-1) - dimensional manifold. Why the -1? Well that's because one of the variables is determined by the equation, so in the case of a line x + 2y = 0, a given x completely determines y.
It is important to notice that I can multiply this equation by any number to produce the same line, e.g. x + 2y = 2x + 4y = 10x + 20y = 0 etc. Any other linear equation will produce a different line.
If you just work with just one of the equations you will never find numerical values for x or y unless it is a horizontal or vertical line. That is because there are an infinite number of solutions (x,y) to a line.
If you are hoping to twiddle 3x + 4y = 24 around to get 4x + 3y = blah, I'm afraid that cannot ever happen. The first equation should be thought of as an assumption/assertion that "this line exists.". It should be obvious intuitively that the second equation, equivalent to the assertion that "this other line also exists.", does not follow logically. A single line existing by itself is perfectly fine, so the existence of a second equation is an assumption plucked out of thin air. You could choose to have more lines if you wanted.
Algebraic twiddling is equivalent to a pure logical argument; which is to say it does not change the properties of the line e.g. 3x + 4y = 24 is the same line as -3x - 4y = -24.
What is special is that we are using (x,y) in both equations, so implicitly we have the added assumption that "there exists a point (x,y) on both lines." Such a point does not always exist; for example if the second equation were 3x + 4y = 25 the second line would be parallel and offset. If there are more than 2 equations of a line in 2D space, it might also fail.
To attempt to find a solution, you simply need to combine the two (or more) equations into a single equation. It really does not matter how you do this, provided you don't stray from the laws of algebra. A common way is to manipulate the variables in both equations independently until you get A = blah and B = blah, then take A = B and solve for x or y. This may not be numerical, but if you substitute your solution for into either of the equations you are certain to obtain a numerical solution.
The problem of solving simultaneous linear equations is extremely important in every technical field and it is right at the heart of quantum mechanics. We call it linear algebra and usually use matrix representations, which are grids of numbers or variables. With this approach you can solve such equations very easily, although the algorithm seems weird until you get a bit deeper into the maths.