Developer Style Guide

R Style Guide

This document describes the coding style used within the package. Having a consistent style enhances the readability and “understand-ability” of the code and makes it easier for users and developers to work with this package and with other, related Mazama Science packages.

Naming Objects

Naming variables is one of the most important things to get right to make your code readable and understandable to future readers of the code (perhaps even yourself!). Having a system for creating names also makes it easier to come up with new ones.

Mazama Science packages embrace lowerCamelCase for object names.

With the casing settled, we use an ornithologist’s sensibility for how to identify things:

What is it? — a bird
What kind of bird is it? — a blackBird
What kind of blackBird is it? — a redwingedBlackBird

It’s a simple system: start with a noun and prefix it with descriptors until it is uniquely identified.

In this system we would never have a variable called: num_hours. Instead we go through our process:

What is it? — (Hmm. What noun describes this? Ah yes!) — a count
What kind of count is it? — (It’s not a “head count” or a “body count”.) It’s an hourCount.

For complex objects it is often helpful to give readers of the code a hint as to what type of object it is so they will know how to work with it. We often use variable names like:

monitor — a mts_monitor object
dataList – a list of dataframes

We occasionally use ’_’ to create classes of similar variables that are otherwise hard to name, e.g.:

QC_negativeValues

Naming Functions

Most functions should strive to be atomic in nature and should do one thing really well. Think of them as individual Lego bricks that we click together to achieve more advanced functionality. Where objects are well described nouns, functions are well described verbs that describe what they do as in:

pas_createNew()
pas_filter()
pas_filterArea()
...

All of these functions begin with pas_ because they are for creating or working with pas objects. Many of these functions accept a pas object as their first argument and return a modified pas. This means that they can be used with the %>% “pipe” operator and chained together as in:

Okanogan_pas <-
  example_pas %>%
  pas_filter(stateCode == "WA") %>%
  pas_filter(countyName %in% c("Okanogan"))

Naming Files

Each file should contain a single function of the same name. Thus, the function named pas_filterArea() is defined in pas_filterArea.R. An exception is made for small, mostly internal functions used in conjunction with a particular type of object or activity. These can be stored together in a file named utils-~:

utils-plot.R

Syntax

We generally adhere to the Wickham Style Guide for syntax with a few exceptions:

Spacing

Do place spaces around code in parentheses if it is an if test:

if ( <logical expression part1> && <logical expression part2> ) {
  ...
}

When debugging, this makes it much easier to select the logical test with a cursor and paste it into the RStudio console.

Lists

We generally like to specify R lists with each parameter = value pair on a separate line. This goes for regular lists and for named argument lists passed to a function:

  pas_createNew(
    countryCodes = "US",
    stateCodes = "WA",
    lookbackDays = 1,
    location_type = 0
  )

Coding this way makes it easy to see which function arguments are being passed. It also eases future refactoring of the code when arguments need to be added or commented out or when the order of arguments need to be changed.

It is our belief that good code should be both readable and understandable and should inspire others to copy and innovate on their own.

Jonathan Callahan