Vjeux

Laurent Senta had the opportunity to go to the 5th European Lisp Symposium to present Climb, the project I've been working on during the past 2 years. He did an excellent job at writing a 4-page paper that sums up the interesting parts of the project (Download).

Presentation

I recommend reading the short article before getting to the slides. Download the PPTX if you want to see the speaker text.

This is the third (and last) presentation about my work on Climb at the LRDE. During the first one I tackled genericity on data structures, the second was about genericity on values and this one talks about genericity on algorithms.

Climb - Property-based dispatch in functional languages

Abstract: "Climb is a generic image processing library. A generic algorithm interface often requires several different specialized implementations. Olena, a C++ library, solves this using properties.

We present a way to dispatch a function call to the best specialized implementation using properties in a dynamic programming language: Common Lisp. Then, we introduce examples of algorithms and properties used in image processing."

In the Javascript world, it is a common thing to chain methods call. For example, this could be a call from an image processing library.

Image('in.png')
  .resize(200, 100)
  .erode()
  .save('out.jpg');

In Lisp, there is not such thing as a dot notation to call an object method. Methods are functions taking the object as first argument. To mimic the dot operator that allows chaining we would like to write:

($ (image "in.png") ; Note: . is not a valid name, we use $ instead
   (resize 200 100)
   (erode)
   (save "out.jpg"))

Hopefully, Lisp allows to rewrite the previous snippet into code that actually works with macros.

With temporary variables

The first way to rewrite it is with a serie of assignement. It uses a temporary variable that is being passed along. progn is being used to group the actions into a single block that returns the tmp value.

(progn
  (defvar tmp (image "in.png"))
  (setf tmp (resize tmp 200 100))
  (setf tmp (erode tmp))
  (setf tmp (save tmp "out.jpg"))
  tmp)

And this is the macro that makes it work.

(defmacro $ (object &rest actions)
  (let ((curr-object (gensym)))
    (concatenate
     'list
     '(progn)
     (list `(defvar ,curr-object ,object))
     (loop for action in actions collect
           `(setf ,curr-object
                  (,(car action) ,curr-object ,@(cdr action))))
     (list `,curr-object))))

Some keys to understand it if you don't know lisp macros.

• ` set the following as output code
• , evaluate the code
• @ expand the list. (resize @(200 100)) -> (resize 200 100)
• gensym creates a local variable with a unique name
• car is the first element of the list, cdr is the rest

Inline

The previous way was probably how would have written it in your code. Since we are programmaticaly rewriting the operation, we do not care about how readable the output is. We can remove the use of the temporary variable inlining the calls.

(save (erode (resize (image "in.png") 200 100)) "out.jpg")

The macros that powers it is much smaller.

(defmacro $ (object &rest actions)
  (let ((res `,object))
    (loop for action in actions do
          (setf res `(,(car action) ,res ,@(cdr action))))
    res))

Conclusion

I took a popular design pattern on the Javascript world and adapted it to lisp. It makes writing several chained method calls easier.