Document uses

[ctbrown]

An interesting use of backpipe is when it is used to clear some clutter. Two uses come to mind.

• Defining S4 Methods : reduces clutter associated with bracketings
• Using logging : Put the logging command first so that it is easy to see.

This makes them more like the way that if works.

#' ## S4 Methods
setMethod( "cx", signature("character","character"), . ) %:% 
  function(x,y) { print("hw") } 

#' ## Logging
 warn %:% "This is not a love song"

There are some variants of this. For example, you could backwardly create a list of arguments that are then passed to the first argument ... always a function.

warn %:% "you've been warned" %:% immediate=TRUE 

Even better if we can give it a special operator:

warn : "you've been warned" 

[mutahiwachira]

Many years later, but did you ever consider using the backpipe to provide syntax matching the mathematical style for 'where'? As in, "let y = f(x) where f(x) is the sine function"

At the time of writing this comment, the backpipe is %<%. So look at this:

`%where%` <- `%<%`

x <- 5
z <- 
  {cube(x)} %where%
  {cube = \(x) x*x*x}

Read it as, "Let x = 5. z is the cube of, where cube means xxx"

I am experimenting with some very heavy use of anonymous functions. I find it weird to read code where I first define the anonymous function and give it a name and then call it. It adds mental overhead and I think it is much more reasonable to say

result <- nicely_named_op(input) %where%
  nicely_named_op = \(x) # ugly set of function calls

I would really like to be able to write

x_turning_point = min_point(parabola) %where%
  {min_point = \(a,b,c) -b/2*a}

Not that this would be useful or good for production code necessarily. But it could be fun.