M05: Structures

The idea of compound data has been around for a very long time. Our presentation of compound data in CS135 and the language we use to describe it is heavily influenced by Object-Oriented Programming, a programming paradigm you’ll learn more about in CS246 (typically a 2A course).

This module was completely rewritten for Fall 2024, so the commentary is still a little sparse :(

Here’s an example of the substitution rules in action.

We need to figure out the substitution rules for structures. This slide formalizes what we’ve said previously to prepare for the rules on the next slide.

The define-struct is followed by a data definition that says what types each of the structure’s fields are expected to be. The data definition for a structure is of the form

;; A <type-name> is a (make-<sname> <type> ...)

where you fill in the parts between angle brackets. <sname> is the first argument to define-struct.

The relationship between the define-struct and its data definition is similar to a function and its contract. Just as you wouldn’t dream of writing a function without a contract, you shouldn’t have a structure definition without a data definition.

The data definition is just a comment, so it has no meaning to Racket. Therefore, Racket does not know what types we expect each field to have. It can’t check, for example, that the description is a string and that prices and available should be numbers.

The fields are listed in the same order in the template as in the define-struct so a new structure can be easily constructed with them.

Students who have programmed in an imperative language will often ask how to change the contents of a field. The answer is that you don’t. You create a new structure with the same information except for the field(s) you want to change.