I would stay away from templates just yet. Keep it as simple as possible. Your class needs to have only two double variables: the real and the imaginary parts.
rootsline1, yes, basically, you are on a right track. However, your present code contains a dozen of syntax errors like c6(5,6,). Try to get rid of them first.
It is much better to use a template to avoid hardcored data type. You are using float one. Right? However, it is not the only choice. A complex number might compose integer or double precision real.
If you're still needing help on this topic I suggest you stop erasing the history of your posts. No one else can jump in to help you now since we can't see the original question.