Interesting preprint posted to arxiv by Carl Boettiger from UC Santa Cruz, describing an example of an increasing trend to use Docker to package up investigations as reproducible research:

Carl Boettiger, An introduction to Docker for reproducible research, with examples from the R environment