Thanks for sharing. Most of the items on this list should be on any organization’s mind! The virtue of using version control is often forgotten I think.

Did you reject other systems, in favor of ggplot2, because they failed to meet some of these criteria? I’m thinking especially of lattice, which is probably more opaque and less popular, but I think does the basic job as well.

The big drawback I see for all the libraries in R is that the font handling is very limited. Accessing non postscript fonts is difficult, and there are no decent typographic controls (even things as basic as letterspacing).

It’s also difficult to really make ggplot2 (or lattice) output conform to a grid system that might exist outside of it (that has been defined for page or web layouts for instance).

(More thoughts on my frustrations with existing DataViz tools here.)

I can see how these issues would not concern a more scientific outfit, where design and visual appeal is secondary. I’m curious if you’ve needed to work around these issues, or whether have there been other limitations with ggplto2 you’ve just decided to live with.