My research explores the utility of natural language descriptions to support a visual language designed for novice programmers. What are the effects of a second representation on program creation? To what extent does it affect the programming processes? Is one secondary representation more effective than another? These initial questions lead this research.