pipeline terminology

[mcascone]

A thought on the terminology you're using. I would tend to think of the whole mini-etl process as a single pipeline, composed of one or more stages, which in turn are composed of one or more steps. In contrast with your terminology of each input>transform>output process being called a pipeline.

This is my Jenkins-colored worldview speaking; the terms could be slightly different, but the point remains: isn't the whole repo an "etl pipeline"? And following, each directory in the repo - each with its own input, output, and transformer - are steps/stages in the flow of the pipeline?