This impromptu talk describes challenges and progress in developing layouts for three different areas. The first very specific area develops linked micromap templates to show county data for different states. Developing a one page design for each state involves special considerations motivated by state shape, county shapes, and county numerosity. (Texas is hopeless.) The second area is somewhat more general and concerns developing graphics to augment recursive partitioning software that has been used on numerous genetic and combinatorial chemistry problems. (Progress toward an example with over two million explanatory variables is slow.) The third area is very general. It concerns layouts for cases, variables, tree models, and other objects where similarity can be computed. Something better needs to be done than principal components, classical multidimensional scaling, and self-organizing maps.

Work is in progress in all three areas. Quite possibly members of the audience will have some good suggestions and students may find a thesis topic.