I guess one way to articulate what you are doing is building /up/ more commands in a step by step fashion with immediate feedback. It is not necessarily a bad thing, especially if processing or preparation times for the input are long. However, you could also accomplish the same thing with a smaller sample size and repeatedly applying your commands using a regular shell instead of an interactive less.