Thanks to a for the replies..... I will study them later when I am free. On Aug 25, 2012 2:37 PM, "nicolas.o...@gmail.com" <nicolas.o...@gmail.com> wrote: > > Here's my take: > > We want to define a function my-comp. It takes n functions and return > their composition. > We want to return a function of any number of arguments, so let's > start by working with a given > set of argument args, and returning the value of the composition > applied to those arguments. > > - If I compose only one function, you get the value of the function itself: > (value-my-comp f) = (apply f args) > > - By definition of composition: > (value-my-comp f1 f2 .... fn) = (f1 (value-my-comp f2 .... fn)) > (Take all the arguments, apply it to the composition of function f2 > .... fn, and then apply the result to f1). > > Then you remove the dots: > > (value-my-comp f1 & f2-fn) = (f1 (apply value-my-comp f2-fn)) > > Then you can just put them all together: > > > (fn [& fs] ;; we take a bunch of functions in > (fn [ & args ] ;; we take the arguments of the resulting function in. > (letfn [(value-my-comp ;; the value of the result in term of the > functions is: > ([f] (apply f args)) ;; only one function > ([f1 & f2-fn] (f1 (apply value-my-comp f2-fn))))] ;; > more than one function > (apply value-my-comp fs)))) ;; we compute the value of the > actual function we got. > > If you don't like (apply value-my-comp f2-fn), you can make > value-my-comp actually takes a sequence > of functions: > > (fn [& fs] ;; take the functions > (fn [ & args ] ;; the arguments of the resulting functions > (letfn [(value-my-comp > [[f1 & f2-fn]] ;; a list of functions > (if (empty? f2-fn) ;; only one function? > (apply f1 args) ;; then we apply the args to it > (f1 (value-my-comp f2-fn))))] ;; else we compose > (value-my-comp fs)))) > This one is another way of writing the one you posted. > (Using a let instead of applying it directly, and not using a > non-varying argument a.) > > If you want to see how this function can be written to be very > performant look at > comp source code, it is very interesting. > > Best, > > Nicolas. > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clojure@googlegroups.com > Note that posts from new members are moderated - please be patient with your first post. > To unsubscribe from this group, send email to > clojure+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en
-- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
