Characterizing Pharo Code: A Technical Report

Pharo (https://pharo.org/) is a modern dynamically-typed reflective pure object-oriented language. It is inspired from Smalltalk. Its unconventional syntax mimics natural language: arguments are not grouped around parentheses at the end but within the message, making expressions looks like sentences. In addition, all control flow operations are expressed as messages and the programmer can freely define new ones and as such define Domain Specific Languages for his task. In this technical report we discuss the statistical properties of source code that people write using Pharo programming language. We present the methodology and tooling for analysing source code selected from the projects of Pharo ecosystem. By analysing 50 projects, consisting of 824 packages, 13,935 classes, and 151,717 methods, we answer questions such as "what is a typical method length?" or "what percent of source code are literals?".