Dr. Dobb's is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Channels ▼


It's Time to Get Good at Functional Programming

Scala: A Hybrid Language

Even if you've used a functional programming language, say in college, unless you've done functional programming on the job to get real work done, you're likely to find that your experience in imperative object-oriented programming will lead you astray in embracing a functional programming mindset. The learning curve is steep.

So what's the best language for exploring FP's benefits? It depends.

Let's say you're heavily invested in Java skills and tools, and don't want to toss aside the code libraries, the skills, and tools you count on. Then you should take a look at Scala.

Scala is a multi-paradigm language that integrates essential features of iterative object-oriented programming with the FP paradigm. It compiles to JVM bytecodes—you can use Java libraries in your Scala programs and inherit from Java classes. You depend on Eclipse? There's a plug-in. Scala lets you isolate those parts of your code that truly need the benefits of FP and write everything else in Java.

Although there is a .NET version, Scala is really married to Java, having been developed by Martin Obersky, one of the designers of Java generics and the author of the current javac reference compiler. It's the Java route to FP. A good Scala site is www.scala-lang.org.

F#: A Functional Language for .NET

Or let's say you're a .NET developer, live and breath .NET, go to all the Microsoft tech conferences, and want to explore FP within the security and convenience of all your familiar tools. You'll want to investigate Microsoft's F#, recently released from the labs and rapidly transitioning to a fully-supported first-class .NET citizen. Like Scala, F# is an OOP/FP hybrid that lets you slip into FP code where you need it and stick with familiar approaches elsewhere.

F# was designed specifically for .NET based on the OCaml FP language; Don Sype led the development in the Microsoft Research Group. It gives your FP code easy access to .NET libraries and tools and is integrated well with Visual Studio.

Both Scala and F# have all the defining features of FP languages: Intense use of pattern matching, functions as first-class values, lazy evaluation. Each gives your code that immutability and statelessness that you need to exploit the multicore opportunity.

Erlang: A No-Compromises Approach

Scala and F# have the virtue of combining OOP and FP paradigms and letting you wade into the FP waters only as deeply as you need to go. But what if you want to dive right in? Then you may want to explore Erlang.

Erlang is a general-purpose language and runtime environment specifically designed by Joe Armstrong at Ericsson for building highly parallel, distributed, fault-tolerant systems. Joe, who might be prejudiced, seems to thinks it's the next Ruby, and claims that if you write an application in Erlang to run on a single-core processor, it will run four times as fast on a 4-core processor without any modification in the code. And he presents data to (almost) confirm that.

Erlang is not a hybrid like Scala and F#, but a no-compromises pure FP language. It does not have "a C-like syntax to make it easy to learn." And you won't find it all that easy to learn. One Erlang programmer who claims to love the language has nevertheless blogged his frustration at the fact that Erlang has three different expression terminators that are context-dependent.

But for the applications for which it was designed, Erlang is delivering the goods. Ericsson uses it for telecom systems work like controlling a switch or telecom apps, database apps, or Internet apps. Ericsson's AXD301 project, a couple million lines of Erlang code, has achieved 99.9999999% reliability. How? "No shared state and a sophisticated error-recovery model," Joe says.

Haskell: A Foundation for Research

Finally, what if you just want to learn this different way of programming? You understand that to get really good at thinking in FP mode you'll need to go back to school (figuratively). You want a pure (not hybrid) FP experience and you aren't expecting to get productive work out of the exploration. Is there a best language for simply learning functional programming?

A good case can be made for Haskell.

At a meeting in Portland in 1987, a group of programmers decided that the FP landscape was getting littered with divergent approaches and minor, dead-end implementations, and that it was necessary to form a committee to design a common functional programming language with the goal of having a single foundation for FP research, education, and promotion of functional programming. That language was Haskell, so yes, Haskell is a camel (a language designed by a committee), and it is also a good language for learning all the FP essentials and the style of thinking needed for functional programming.

Haskell is a no-compromises functional programming language like Erlang. It presently has no commercial implementations. But while its aspirations are more academic than commercial, there are some interesting applications of Haskell. Linspire uses Haskell for system tools development, for example, and XMonad, a window manager for the X Window System, is written entirely in Haskell. Haskell's influence can be seen in the LINQ features of C# 3.0 and elsewhere.

So you have some choices, and I haven't even mentioned Intel's language Ct or Caml or Mathematica (but see the sidebar). The sharpest distinction among FP languages is between the hybrid approaches that let you extend your current skillsets and toolkits to encompass FP and the pure FP languages that require you to move fully into their world, at least for a while. Probably both approaches are necessary, and it will be interesting to watch how this all plays out. Because one way or another, functional programming is on the rise.

— Michael Swaine

Functional Programming in Mathematica

by Mike Riley

In addition to being a rich mathematical expressions parser, Wolfram Research's Mathematica delivers a complete programming environment. Mathematica's programming model is extremely flexible, letting you create math-centric applications using standard procedural, object-oriented, and functional programming approaches. As Wolfram's Jon McLoone explains, "In the most simplistic sense, we can consider functional programming to be when we pass functions as arguments into other functions" (www.wolfram.com/broadcast/screencasts/elementaryprogramming). He then goes on to show how easy it is to leverage the utility of this powerful programming approach. In McLoone's example, a compound accumulation of interest is calculated in two lines of code. The first defines an addinterest function using Mathematica's function syntax:

addinterest[n_]:= n*1.04

The syntax is straightforward: addinterest is the name of the function, the brackets contain the parameter(s) to be passed into the function, the := is Mathematica's function-assignment operator, followed by the actual function itself. The second wraps this addinterest into a defined Mathematica nested list function called, appropriately enough, NestList:

NestList[addinterest, 100, 20]

This generates the accumulation of 20 compound interest results starting with the capital amount of 100. Using this facility, highly sophisticated functional relationships can be constructed in minutes.

Once the essentials of functional programming using Mathematica are grasped, using Mathematica as a primary programming environment is a natural progression. At the International Mathematica User Conference (www.wolfram.com/news/events/userconf2008), I interviewed several professors, engineers, and scientists on their use of the program, all of which were strong proponents of Mathematica's functional programming capabilities. For example, mathematician Eric Schulz said that "over the years, I've coded in Fortran, I've coded a little in C, I've coded in Perl. I've done quite a bit of Visual Basic in the Windows world. My favorite language by far is the Mathematica language as a programming environment." Schulz's own programming work will be prominently featured in one of the "assistants" in the upcoming release of Mathematica 7. Assistants are a new extension capability that let Mathematica developers enhance the application's environment with custom features. Schulz's assistant exposes numerous visual constructs and functions in an easily accessible palette, and was written to scratch his own itch of visually interacting with Mathematica when teaching students using SmartBoard (smarttech.com).

At the same conference, Continuity Logic's Kris Carlson demonstrated conference a simulation he wrote that modeled some of the signaling pathways that govern aging. "I could program in Perl, Visual Basic, and SQL, and I was getting interested in Python. Frankly, I was a jack-of-all-trades...but I decided to drop those and focus just on Mathematica...If I had to go back to those other languages, I would need psychotherapy because I really believe Mathematica is such a superior language."

For an engaging conversation with Wolfram Research's founder Stephen Wolfram, check out Dr. Dobb's TV).

Mike is a DDJ contributing editor. He can be contacted at [email protected]

Related Reading

More Insights

Currently we allow the following HTML tags in comments:

Single tags

These tags can be used alone and don't need an ending tag.

<br> Defines a single line break

<hr> Defines a horizontal line

Matching tags

These require an ending tag - e.g. <i>italic text</i>

<a> Defines an anchor

<b> Defines bold text

<big> Defines big text

<blockquote> Defines a long quotation

<caption> Defines a table caption

<cite> Defines a citation

<code> Defines computer code text

<em> Defines emphasized text

<fieldset> Defines a border around elements in a form

<h1> This is heading 1

<h2> This is heading 2

<h3> This is heading 3

<h4> This is heading 4

<h5> This is heading 5

<h6> This is heading 6

<i> Defines italic text

<p> Defines a paragraph

<pre> Defines preformatted text

<q> Defines a short quotation

<samp> Defines sample computer code text

<small> Defines small text

<span> Defines a section in a document

<s> Defines strikethrough text

<strike> Defines strikethrough text

<strong> Defines strong text

<sub> Defines subscripted text

<sup> Defines superscripted text

<u> Defines underlined text

Dr. Dobb's encourages readers to engage in spirited, healthy debate, including taking us to task. However, Dr. Dobb's moderates all comments posted to our site, and reserves the right to modify or remove any content that it determines to be derogatory, offensive, inflammatory, vulgar, irrelevant/off-topic, racist or obvious marketing or spam. Dr. Dobb's further reserves the right to disable the profile of any commenter participating in said activities.

Disqus Tips To upload an avatar photo, first complete your Disqus profile. | View the list of supported HTML tags you can use to style comments. | Please read our commenting policy.