Clojure weekly, Feb 05th, 2013
Welcome to another issue of Clojure weekly, my small routine blog contribution to the Clojure sphere! These are just a few links, normally 4/5 urls, pointing at articles, documentation, screencasts, podcasts or anything else that attracts my attention. I add a small comment so you can decide if you want to look at the whole thing or not. That’s it, enjoy! Reducers - Rich Hickey on Vimeo Here's some hardcore stuff for well-rounded clojurian. Well, Reducers are coming in 1.5 (RC4 at the time of this writing) so you should know how you can use them and for which kind of problems as everything else in the standard library. Rich the-man is going to take you through reducers in this talk from Euroclojure. The content is certainly dense and some background material is recommended at the beginning. On the understandable side, Rich is giving code examples, from the basic form of a reducing transformer to its final form. It is highly recommended you fire up a REPL and follow along, or even better, stop and create a sample project. Reducers are essentially a trick, a combination of reduce over map (or filter) where the mapping operations can be accumulated until reduce will eventually realise. The trick has the nice side effect to remove ordering from standard map operations, making them parallelizable. The new fold function based on reducers doesn't use any parallel-map or parallel-collection specific stuff, an object oriented aberration. Reducers without the fold context aren't that different from monads or combinators: they are functional patterns. Somebody should write a book about it. ClojureDocs - clojure.core/mapcat Mapcat is somehow an useful shortcut. It is mostly used whenever a map operation returns a collection of collections. This happens when the f passed to map is returning a collection of some sort. To flatten that nested collections, you could use concat over map returned list like (apply concat [[1], [2]]). Or just enter mapcat that does the same thing for you. By using mapcat you agree that the mapping function passed to map is returning collection instead of scalars, otherwise the mapcat operation will just fail. Clojure - sequences This is one of the those topics somebody looking at Clojure for the first time will probably come across. Sequences are an abstraction on collections which contains their notion of ordering in terms of what is "first" and what comes "next". You can "first" on a sequence, you can "rest" to access everything after the first element or you can "cons" an item and a sequence to create a new (cell) sequence. There is a reason why they are particularly important: Clojure laziness is build on top of them. Almost all sequences (collections that are seq-able) are also lazy which basically boils down to the way "rest" is implemented: it returns a function which is a recipe on how to realise the rest of the collection instead of a concrete collection. Agile Toolkit Podcast - Arlo Belshee I've found an interesting take of Arlo Belshee about Midje mocks in this episode of the Agile Toolkit Podcast. It is toward the end, but the podcast is short, I suggest you to listen to the whole thing especially if you don't know Arlo. For state based paradigms like OOP state is what behaviour gravitate around. When mocks are used in OOP you are mostly mocking state, so that you can test the behaviour given a particoular context. But with functional programming, state is spread everywhere and in a sense it is state that is sent around to be handled by behaviour. Arlo doesn't like mocking in OOP preferring state based testing in the Kent Beck sense for this reason, because mocking breaks the design flow. But mocking in FP is really about behaviour: injecting new behaviour to verify how state changes, exactly like state based testing in OOP. This is why Arlo likes it.
