If you mean "rearranging a quadratic so you can read the vertex off the equation", then the basic process is as follows:
. . . . .y = ax<sup>2</sup> + bx + c
. . . . .Factor out whatever is multiplied on the squared term,
. . . . .but factor it only from the variable terms.
. . . . .y = a[x<sup>2</sup> + (b/a)x] + c
. . . . .Take half of whatever is now multiplied on the linear
. . . . .("x") term, and square it. To "complete the square",
. . . . .you'll be adding this inside the brackets; to keep that
. . . . .side "even" (unchanged), you'll also be subtracting it
. . . . .from the same side.
. . . . .y = a[x<sup>2</sup> + (b/a)x + (b/2a)<sup>2</sup>] + c - a[(b/2a)<sup>2</sup>]
. . . . .Now convert to "complete the square" form:
. . . . .y = a[x + (b/2a)]<sup>2</sup> + c - (b<sup>2</sup>/4a)
Simplify, and read off the values for h and k.
Eliz.