1 Idioms

1.1 The four basic principles to a tidy API

These principles are likely to change. This version is from January 1st 2016.

1.1.1 Reuse existing data structures.

• try to use existing common base R structures instead of new customized ones
• if the latter is not possible create S3 classes based on a base R atomic or list
• use data frames or tibbles for rectangular data made up of observations and variables in rows and columns (which is basically tidy data)
• for “non-standard scoping”, prefer formulas over non-standard evaluation and use the lazyeval package.

1.1.2 Compose simple functions with the pipe.

1.1.3 Embrace functional programming.

1.1.4 Design for humans.

1.2 Resources

• The tidy tools manifesto