As domain-specific languages (DSLs) permeate into mainstream software engineering, there is a need for economic methods for implementing languages. Following up on a paper with a similar title, this paper focuses on dynamically typed languages, covering issues ranging from parsing to defining and staging an interpreter for an interesting subset of Dr. Scheme. Preliminary experimental results indicate that the speedups reported in previous work for smaller languages and with smaller benchmarks are maintained. © 2008 Springer Berlin Heidelberg.