My Own Style in R

Although I have written programs since I were a high school student, when computer monitor could only display one color, I have no formal training in programming, and I rarely worked with others in developing a solution until recently. The problem: I did not write with a consistent and professional style. I am pretty sure that my code will look “ugly” to professional programmers.

That said, I do have a loose style, one that suits my own situation:

• I work on small screen frequently, sometimes even on my mobile phone. I wrote the documentation of some packages on my mobile phone, and even this post was largely drafted on my phone.

• My main tasks are research and teaching, among other tasks. I can easily forget some style rules I set for coding. I need something simple and easy to remember.

• Ideally, code should be easy, or at least not too difficult, to comprehend by future me with minimal comments. What looks nature to me myself is of the top priority.

So, this is my style, with me as the main user and reader:

• For code, I use a line width of 60 to 70. For documentation, I am more aggressive and use a line width of 40.

• For the same reason, I use two spaces for indentation. A four-space indentation is too “expensive” to me.

tmpfct <- function(x) {
    x^2
  }

• I like using double-click to select a name. This does not work if periods are used inside a name. Therefore, I no longer use periods in a name, except for S3 methods. If necessary, I use underscores.

this_is_long_name <- 1

• I found it difficult to remember the case I used for a name. So, I stick to lowercase letters unless I am very certain that I can remember that I used uppercase letters.

I don’t like camel case. It is OK for language that is not case sensitive, like Visual Basic and SPSS syntax commands, but is inconvenient for case sensitive languages like R.

# I don't like camel case.
thisIsNotWhatIDo <- 1
# I prefer this:
this_is_what_i_do <- 1

• Long function names are acceptable with me. With autocompletion in many IDEs, it is not important to use short names. Being easy to remember part of a name is important. An abbreviation is not easy to remember unless it is commonly used (e.g., SD).

# This is easy to remember
factor_loadings()
# These variants are not
fload()
facload()
facload()
fl()

• I wrote stuff in Python occasionally. I like the Python style indentation, which is easy to read. So I use that style for my R code too.

for (j in 1:10) {
    # Do something
  }
if (x == 1) {
    # Do something
  } else {
    # Do something else
  }

• I never, ever, use any automatic stylers to reformat code. They make changes that are tracked by Git but have nothing to do with the content. I may use them, but only when finalizing the code.

• This is also why I care little about word wrap. Irregular line widths are acceptable for me.

• Extra whitespace are OK with me. Readability is the main goal.

# I may do this:
x  <-   1
y0 <- 100

• I use double quotes for string literals. I have to use two keys … but the habit is too difficult to break that I don’t bother changing it.

• The last “rule”: I can break any of the rules, as long as the code is readable without the need to know any rules.

I also have a GitHub repo for my personal style, in case I forgot the rules:

https://github.com/sfcheung/rstylesf

So, please pardon me if you find my code for packages at odd with professional style. I myself is the main reader and maintainer of the packages. What work for me matters.

:::