The whole point of a parameter in Fortran is that the parameter's value is known at compile time, so the compiler knows how much memory to allocate for the array. For a variable (not a parameter), the variable gets its value at run time, I believe.
Minor point. A statement such as y = 2 is an assignment, or in this case an initialization, not an equation. It does not state that y and 2 are equal. Instead, it stores the constant value 2 in the location that is named y.
Most programming languages make distinctions between assignment and testing for equality by using different symbols. For example, Fortran uses .EQ. to test for equality, and uses = for assignment. Languages based on C, such as C++, C#, Java, and others, use == to test for equality and = for assignment. Newer versions of Fortran now support the C-style notation, in addition to the older .EQ., .LT., GT., and other spelled-out relational operators.