Qn re operators for binary- vs real-variable representation

[kc_@ejazzmail.com]

hi there

    i'll like to clarify the following:

1)    i have come across material that mentions that the choice of
representing one's variables ( say b/w binary-valued or real-valued
variables) would affect one's selection and mutation operators for
effecting GA to solve a particular problem.  To what extent do i have to
take this into consideration when i built my own GA-software?

2) I was intending to apply several types of 'binary mutation operators'
( like single pt and multi pt ) to real-valued variables.  In view of
(1), does that mean such application will be 'wrong'??

    I look forward to any help / insight from similar encounters...TIA

ken