Explanation of the curry function for node.js

node-curry.js

function curriedReadFile(path) {
  var _done, _error, _result;

  var callback = function(error, result) { 
    _done = true,
    _error = error, 
    _result = result; 
  };

  fs.readFile(path, function(e, r) { 
    callback(e, r); 
  });

  // Here '_' is the function we pass to curriedReadFile
  // Where we actually do the IO stuff
  return function(_) {

    // If fs.readFile returned (_done is set), we just execute our IO code with the results
    if (_done) {
      _(_error, _result);

    // If it still hasn't return, our function that does IO stuff ('_') 
    // now becomes the callback (see fs.readFile body)
    } else {
      callback = _; 
    };
}

Is there a reason for wrapping callback(e, r)?

Ie, why not simply:

fs.readFile(path, callback);