This analogy is not quite right. > (fn [n] (vector n (+ n 1))
is not a reducing fn. Tt is a cons cell builder, and the Church encoding builds lists. The point of this library is to define map/filter etc *without* using lists/streams - not as input, not as output, not producing links/thunks etc, nor implicit streams formed by first/rest chained recursion, nor presuming input collections defined similarly. It is only by doing this that you can get the true prize of this library - versions of map/filter etc that can be run in parallel, and that are completely independent of the data structures on which they are run. Rich On May 9, 2012, at 7:09 AM, nicolas.o...@gmail.com wrote: > It looks very good. > > Just a few side-notes: > - this representation is quite well known, as it is the Church > encoding of list, often used in System F, for example. > It corresponds to stating the definition of sequences as the > initial algebra of F A = 1 + Any x A. > - you can play a dual trick for Streams and coLists by representing > them as how to generate them, the same way that you can represent list > as how you reduce them. > This gives a representation of a coList as > [ seed f ] where f is a function from the seed to either nil(if the > coList is finished) or a pair of an element and as seed. > For example, the integers would be > nat = [ 0 (fn [n] (vector n (+ n 1))) ] > Then, you can play the same game and define map, filter (which is not > always productive...), and so on, on this > representation, with the same benefits. > > You can then also write converters from this representation to > reducible. This would allow to write > (reduce + (take n nat)) without allocating any sequence... > > 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
