The idea here is to come up with a visualization of source code that would give some indication of quality to non-programmers.
Imagine for example a service that indexes open source code and provides such a visualization in the form of gardens. A non-programmer could quickly examine and compare the gardens for some competing open source software applications.
Bad smells in the source code would manifest themselves in the form of a poorly maintained garden.
You could also show code evolution via an animation of the garden over time.