scopt is a little command line options parsing library.

Today, I'm releasing scopt 3.0. If you're not interested in the implementation details, skip to the readme.

Around March 4th, 2010, I became a committer to scopt, a fork of Aaron Harnly's scala-options that was written in 2008. I think I wanted to make a few changes around the usage text, key=value options, and argument list. Since then I've been fielding all the bug reports, including the request to publish the jar on scala-tools.org. On March 18, 2012, I forked the project again to scopt/scopt and released scopt 2.0.0 that added immutable parser.

After years of adding features on top of the other, I decided to rewrite scopt3 from scratch. The tipping point was Leif Wickland asking if there's a "philosophical reason that scopt doesn't have an intArg().".

Dispatch has been the de facto library to get to the Internet from Scala. To keep in step with the recent move towards non-blocking IO, @n8han rewrote the library as Reboot based on Async Http Client. This became Dispatch 0.9. Then Dispatch 0.10 came out to replace its own Promise type with the standarized SIP-14 Future.

As with Dispatch Classic, Reboot lets you write plugins to wrap around web APIs. In this post we'll port a plugin from Classic to explore how it works.

This is a translation of 「モナドはメタファーではない」に関する補足 by Kenji Yoshida (@xuwei_k), one of the most active Scala bloggers in Japan covering wide range of topics from Play to Scalaz.

Daniel Spiewak's Monads Are Not Metaphors was written about two and a half years ago, but seeing how its Japanese translation is still being tweeted and being bookmarked by 250 users on Hantena, its popularity doesn't seem to cease. I just remembered something to note about the example code used in the post, which could be an unstylish critique, but I'm going to jot it down here. It's an unstylish critique, because I'll be digging into the part where the author likely knew from the beginning but omitted it intentionally for the sake of illustration. Also I'll be using Scalaz in this post.

Scalaz 7.0 is out. For easier reading and toc, I converted learning Scalaz series into an html5 book using @n8han's Pamflet:

• learning Scalaz

It's been a month now, but on March 1, 2013 I flew to Japan to attend "Scala Conference in Japan 2013." That's the name of the conference.

from a podcast

One day (June 2, 2012), I was listening to Scala Types recorded at Scala Days 2012, and someone (turns out it's @timperrett) said "I would love to see Scala Days in Asia. We had two in Europe now. It would be wicked to have it in China or Japan, or somewhere like that." I relayed this info on twitter saying there's an opportunity for the Scala community in Japan to invite the conference.

This is the 5th entry of Scalaz Advent Calendar 2012.

During the months of December, tech-savvy geeks in Japan take turns to post themed blog articles, known as the "Advent Calendar". For last year's Scala Advent Calendar 2011 I translated Eric Torreborre's post covering The Essence of Iterator Pattern. It was somewhat of a calculated move, knowing Japanese fondness for functional programming articles. Another selfish motive was that some of the concept would seep in to my thickness as I was translating the post word by word. In hindsight, both goals were achieved handsomely thanks to the quality of both Jeremy Gibbons, Bruno Oliveira and Eric's work. This seeped in knowledge was probably the secret sauce behind the learning Scalaz series that I worked on this year.

As covered in learning Scalaz day 12 Scalaz 7 already included product and compose methods in typeclass instances as well as Traverse. It even has the word count example from the paper. What I realized missing was the value-level composition. One of the interesting points from the paper was "composition of applicative functors," which enables a kind of modular programming.

By "applicative functors" Gibbons and Oliveira actually mean composition of applicative functions, not just the typeclass instances. This is evident in the following snippet from the paper:

data (m ⊠ n) a = Prod { pfst :: m a, psnd :: n a }
(⊗) :: (Functor m, Functor n) ⇒ (a → m b) → (a → n b) → (a → (m ⊠ n) b)
(f ⊗ g) x = Prod(f x)(gx)

The algebraic data type ⊠ is the type-level product, while the infix function ⊗ is the value-level product of two applicative functions, which returns applicative function of type a → (m ⊠ n). In other words, the programmer would construct functions that return an applicative functor, and the type-level compositions are done automatically.

Hey there. There's an updated html5 book version, if you want.

On day 17 we looked at IO monad as a way of abstracting side effects, and Iteratees as a way of handling streams. And the series ended.

Func

I wanted to continue exploring a better way to compose applicative functions, and came up with a wrapper called AppFunc:

val f = AppFuncU { (x: Int) => x + 1 }
val g = AppFuncU { (x: Int) => List(x, 5) }
(f @&&& g) traverse List(1, 2, 3)

Hey there. There's an updated html5 book version, if you want.

Daniel Steger for openphoto.net

Yesterday we looked at Memo for caching computation results, and ST as a way of encapsulating mutation. Today we'll continue into IO.

IO Monad

Instead of reading the second half of the paper, we can get the gist by reading Towards an Effect System in Scala, Part 2: IO Monad by Rúnar (@runarorama):

While ST gives us guarantees that mutable memory is never shared, it says nothing about reading/writing files, throwing exceptions, opening network sockets, database connections, etc.

Hey there. There's an updated html5 book version, if you want.

Yesterday we looked at Arrow as a way of abstracting function-like things and Unapply as a way of providing typeclass meta-instances. We also continued on with the applicative experiment by implementing XProduct that supports parallel compositions.

Memo

Pure functions don't imply they are computationally cheap. For example, calcuate a list of SHA-1 hash for all permutations of ASCII character string up to 8 characters length. If we don't count the tab character there are 95 printable characters in ASCII, so let's round that up to 100. 100 ^ 8 is 10 ^ 16. Even if we could handle 1000 hashing per second, it takes 10 ^ 13 secs, or 316888 years.

Hey there. There's an updated html5 book version, if you want.

Rodolfo Cartas for openphoto.net

On day 14 we started hacking on Scalaz. First, typeclass instances for Vector was added to import Scalaz._. Next, we rolled back <*> to be infix ap. Finally, we added an implicit converter to unpack A as [α]A, which helps compiler find Applicative[({type λ[α]=Int})#λ].

All three of the pull requests were accepted by the upstream! Here's how to sync up:

$ git co scalaz-seven
$ git pull --rebase