within package variation

[yenchiayi]

• a number for each package. e.g. distance

[yenchiayi]

The preliminary result on the latest submission, after normalizing entropy into [0, 1]

• less consistent: fx_integer, fx_comma, fx_infix, fx_opencurly, fx_name
• more consistent: fx_assign, fx_singleq, fx_closecurly, fx_t_f, fx_semi, fx_tab

• path: visualization_fun/within_pkg_feature_entropy.png

[yenchiayi]

Visualize the naming variation of the 20 most popular packages

• source: https://www.r-pkg.org/downloaded
• it is clear that lower_snake is the dominating naming style, which is far leading the lowerCamel case that used to be the most popular naming style.

• path: visualization_fun/within_pkg_naming_variation.png

[chainsawriot]

Visualize the naming variation of the 20 most popular packages

* source: https://www.r-pkg.org/downloaded

* it is clear that lower_snake is the dominating naming style, which is far leading the lowerCamel case that used to be the most popular naming style.

* [path](https://github.com/chainsawriot/rstyle/blob/entro/visualization_fun/within_pkg_naming_variation.png): visualization_fun/within_pkg_naming_variation.png

@exilespacer Thank so much for this. But the problem of "popular packages" from r-pkg is that nearly all of them are from RStudio (and especially, from a particular team of RStudio. rsconnect is also an RStudio product, but it is from the Shiny team. Thus lowerCamel is used.)

Could we have another chart as a supplement to this, that is, using the top 20 packages with the highest PageRank instead?

[yenchiayi]

@chainsawriot Here is a figure of within-packages naming variation for the packages of the largest pagepank.

• lower_snack / lowerCamel / dotted.func are the 3 most popular naming conventions
• the adoption of naming conventions seems to be related to the time these packages are developed. (dotted.func -> lowerCamel -> lower_snake)
• [most recent] The newly-emerging rstudio-related packages extensively use lower_snake case, which is shown dominating in the most important 20 packages
• [earliest]The packages that exist since the earliest days, such as Lattice or MASS, still stick to the use of dotted.func
• [middle-age] The packages that are widely used but may not be the earliest-appearing packages, such as XML or RUnit, use lowerCamel
• Furthermore, some packages implement inconsistent naming conventions, such as survival and Rcpp, which are noteworthy.

• path: visualization_fun/within_pkg_naming_variation.png

[yenchiayi]

ok, the below figure seems to not support my hypothesis.

I think it will be interesting to look at the relation between the age of packages and the most popular naming convention of the packages. I have a feeling that it will be positively related to the order of (dotted.func -> lowerCamel -> lower_snake)