Executable and Declarative Encoding and Algebraic Design

Executable vs. declarative encoding

Executable encoding is most similar to what we’ve done in previous chapters, while declarative encoding is similar to the two techniques we just discussed. Our DSL is just data that needs interpretation and can have multiple interpreters, and we have the option to define more of them whenever we desire. While the former is simpler, the latter has more power and flexibility.

Algebraic design

Another term worth mentioning is algebraic design. An algebra (mentioned earlier in this chapter) is a group of functions that work with certain types of data, with an accompanying set of laws. We design our application starting with this algebra. Only later do we decide on a representation for it. We create types, which serve as our interface, and later write an implementation for it all. Once again, the idea is to first clearly model our domain. The actual implementation is of no concern yet. In fact, working on it might distract us. This is similar to TDD, where we focus on working on what we want to accomplish through writing tests before we turn to the how.

John A De Goes writes: “Most functional code doesn’t solve problems directly. Instead, it creates models of solutions to problems, which are later executed (or interpreted).”