Let's see why the D programming language is worth a serious look.
Of course, I'm not deluding myself that it's an easy task to convince you. We programmers are a strange bunch in the way we form and keep language preferences. The knee-jerk reaction of a programmer when eyeing a The XYZ Programming Language book on a bookstore shelf is something like, "All right. I'll give myself 30 seconds to find something I don't like about XYZ." Acquiring expertise in a programming language is a long and arduous process, and satisfaction is delayed and uncertain. Trying to find quick reasons to avoid such an endeavor is a survival instinct: the stakes are high and the investment is risky, so having the ability to make a rapid negative decision early in the process can be a huge relief.
That being said, learning and using a programming language can be fun. By and large, coding in a language is fun if the language does a satisfactory job at fulfilling the principles that the coder using it holds in high esteem. Any misalignment causes the programmer to regard the language as, for example, sloppy and insecure or self-righteous and tedious. A language can't possibly fulfill everyone's needs and taste at the same time as many of them are contradictory, so it must carefully commit to a few fundamental coordinates that put it on the landscape of programming languages.
So what's the deal with D? You might have heard of it already -- the language with a name like a pun taken a bit too far; annoyingly mentioned now and then on newsgroups dedicated to other languages before the off-topic police reprimands the guilty; praised by an enthusiastic friend all too often; or simply as the result of an idle online search a la "I bet some loser on this big large Internet defined a language called D, let's see... oh, look!"
In this article I provide a broad overview, which means by necessity I use concepts and features without introducing them rigorously as long as they are reasonably intuitive.
Let's take a brief look at some of D's fundamental features. Be warned that many features or limitations come with qualifications that make their boundaries fuzzy. So if you read something that doesn't quite please you, don't let that bother you too much: the next sentence may contain a redeeming addendum. For example, say you read "D has garbage collection" and you get a familiar frozen chill up the spine that stops in the cerebrum with the imperious command "touch the rabbit foot and stay away." If you are patient, you'll find out that D has constructors and destructors with which you can implement deterministic lifetime of objects.
But Before Getting Into It...
Before getting into the thick of things, there are a few things you should know. First and foremost, if you kind of considered looking into D for whatever reason, this time is not "as good as any," it's in fact much better than others if you're looking for the edge given by early adoption. D has been evolving at a breakneck pace but in relative silence, and a lot of awesome things have been and are being done about it that are starting to become known just about now -- some literally in this very article. At this writing, my book The D Programming Language is 40% complete and available for pre-order at Amazon. Safari's Rough Cuts subscription-based service makes advance chapters available here.
There are two major versions of the language -- D1 and D2. This article focuses on D2 exclusively. D1 is stable (will undergo no other changes but bug fixes), and D2 is a major revision of the language that sacrificed some backwards compatibility for the sake of doing things consistently right, and for adding a few crucial features related to manycores and generic programming. In the process, the language's complexity has increased, which is in fact a good indicator because no language in actual use has ever gotten smaller. Even languages that started with the stated intent to be "small and beautiful" inevitably grew with use. (Yes, even Lisp. Spare me.) Although programmers dream of the idea of small, simple languages, when they wake up they seem to only want more modeling power. D's state of transition is putting yours truly in the unenviable position of dealing with a moving target. I opted for writing an article that ages nicely at the expense of being occasionally frustrating in that it describes features that are in the works or are incompletely implemented.
The official D compiler is available for free off digitalmars.com on major desktop platforms (Windows, Mac, and Linux). Other implementations are underway, notably including an a .NET port and one using the LLVM infrastructure as backend. There are also two essential D libraries, the official -- Phobos, and a community-driven library called Tango. Tango, designed for D1, is being ported to D2, and Phobos (which was frustratingly small and quirky in its D1 iteration) is undergoing major changes and additions to take full advantage of D2's capabilities. (There is, unsurprisingly, an amount of politics and bickering about which library is better, but competition seems to spur both into being as good as they can be.)
Last but definitely not least, two windowing libraries complete the language's offering quite spectacularly. The mature library DWT is a direct port of Java's SWT. A newer development is that the immensely popular Qt Software windowing library has recently released a D binding (in alpha as of this writing). This is no small news as Qt is a great (the best if you listen to the right people) library for developing portable GUI applications. The two libraries fully take D into "the GUIth dimension."