Work in physics is not the same as everyday effort. A man holding a heavy suitcase while standing still gets tired, but in the language of physics he does zero work on the suitcase, because it has no displacement. Work is done only when a force succeeds in moving its point of application.
Work done by a constant force. If a constant force $\vec{F}$ acts on a body and the body undergoes a displacement $\vec{s}$, the work done is the scalar (dot) product $W=\vec{F}\cdot\vec{s}=Fs\cos\theta$, where $\theta$ is the angle between the force and the displacement. Work is a scalar; its SI unit is the joule (J), where $1\ \text{J}=1\ \text{N}\cdot\text{m}$. The sign of work depends entirely on $\cos\theta$:
- Positive work ($0\le\theta<90^\circ$): the force has a component along the motion, e.g. gravity on a falling stone.
- Zero work ($\theta=90^\circ$): the force is perpendicular to motion, e.g. centripetal force on a body in a circle, or the normal reaction on a block sliding horizontally.
- Negative work ($90^\circ<\theta\le180^\circ$): the force opposes motion, e.g. friction, or gravity on a stone thrown upward.
Work done by a variable force. Most real forces change with position. Over a tiny displacement $dx$ the force is nearly constant, so the small work is $dW=F\,dx$. Adding all such bits gives $W=\int_{x_1}^{x_2} F\,dx$. On a force-displacement graph this integral is simply the area under the $F$-$x$ curve. This is the most powerful way to compute work for springs and other position-dependent forces.
Kinetic energy is the energy a body possesses because of its motion. For a body of mass $m$ moving with speed $v$, $KE=\frac{1}{2}mv^2$. It is a scalar, always positive, with the same unit (joule) as work. Since momentum $p=mv$, kinetic energy can also be written as $KE=\frac{p^2}{2m}$.
The work-energy theorem ties work and motion together: the net work done by all forces on a body equals the change in its kinetic energy, $W_{net}=\Delta KE=\frac{1}{2}mv^2-\frac{1}{2}mu^2$. A short derivation for constant acceleration: from $v^2=u^2+2as$ we get $\frac{1}{2}mv^2-\frac{1}{2}mu^2=mas=Fs=W_{net}$. The theorem holds for variable forces too and is often far quicker than using $F=ma$ directly, because it bypasses time and acceleration.