The paper systematically studies neural architecture search for graph neural networks by proposing (1) a general GNN design space, (2) a GNN task space with a quantitative similarity metric and (3) the design space evaluation. Although it is experiment-driven and lacks deep insights or theoretical analysis, the comprehensive and systemic evaluation of GNN design are important to the community. Based on the reviews, the merits of the paper outweigh the drawbacks and acceptance is recommended.