A linear program is an optimization problem of the form
$$\begin{align*} \text{minimize }& \transpose{ \b{c}}\b{x} \br \text{subject to }& \b{A} \b{x} = \b{b} \br & \b{x} \geq \b{0} \end{align*}$$
where $\b{c} \in \R^n, \b{b} \in \R^m,$ and $\b{A} \in \R^{m \times n}$.
This form is called the standard from.
Here $\b{A} $ is an $m \times n $ real matrix, $m < n, \rank{ \b{A}} = m$.
Without loss of generality, we can assume $\b{b} > 0 $, for if $\b{b}_i < 0$, we multiply the $i$th constraint by $-1 $ to obtain a positive $\b{b}_i$.
The constraints may also be in the form of inequalities, such as follows:
$$\begin{align*} &\text{minimize }& \transpose{ \b{c}}\b{x} \br &\text{subject to }& \b{A} \b{x} \geq \b{b} \br && \b{x} \geq \b{0} \end{align*}$$
This is still a linear program, and can be rewritten into the standard form by introducing surplus variables $y_i $ to convert the above into
$$\begin{align*} &\text{minimize }& \transpose{ \b{c}}\b{x} \br &\text{subject to }& \b{A} \b{x} - \b{I}_m \b{y} = \b{b} \br && \b{x} \geq \b{0}, \b{y} \geq \b{0} \end{align*}$$
On the other hand, if the problem is like this:
$$\begin{align*} &\text{minimize }& \transpose{ \b{c}}\b{x} \br &\text{subject to }& \b{A} \b{x} \leq \b{b} \br && \b{x} \geq \b{0} \end{align*}$$
Then we introduce slack variables $ y_i $ to convert the above into
$$\begin{align*} &\text{minimize }& \transpose{ \b{c}}\b{x} \br &\text{subject to }& \b{A} \b{x} + \b{I}_m \b{y} = \b{b} \br && \b{x} \geq \b{0}, \b{y} \geq \b{0} \end{align*}$$