The short answer to "What's the state of Dart?" is: it's in Technology Preview. That's a special way of saying, "we launched early so we can open source everything and work in the open." "Technology preview" also means "we're not even in Alpha yet, we have a lot of work to do, but there's enough there for you to play with and give feedback."
Internet time may work for news stories or consumer product iterations, but probably not for something as ambitious and broad as the Dart effort. Remember, Dart is more than just a language. It's also a set of libraries, a better DOM interface, a virtual machine, an Editor, and integration with Chrome. The team is working very hard on a lot of parallel threads, but I personally expect it'll be six months before we have most of the pieces in place.
The big picture is that Dart will become a "batteries includes" development environment for modern web apps. Dart's driving goal is to help ensure the web remains a productive and enjoyable platform for app development and deployment. This means a lot of pieces need to fall into place: language, libraries, editors, virtual machines, and browser integration. Put all together, we believe Dart will be a compelling option for modern web app developers.
The big big BIG picture is that we want to bring app developers to the web, and we want web developers to write more complex web apps. If they use Dart, that's great. But at the end of the day, the language doesn't matter. The only thing that matters is that complex, client side, high fidelity, low latency, beautiful modern web apps are being built.
The language is in a state of development. We see new releases to the spec approximately once per month. Major features are missing, such as reflection, but we keep iterating. We just added map() support to Collection, for example. Gilad Bracha, a guy who knows his languages (having created NewSpeak and worked on the Java Lang Spec) and Josh Bloch, a guy who knows his libraries (having written Effective Java and worked on the Java Collection libraries) are working on the language and libraries, along with the greater team.
As for market share, it's extremely early in the game. It's probably not the right question to ask, as Dart isn't even shipping. A more interesting question would be, "What is the market share of apps on the web?" and then go figure out how we can address that.
As for a Web App showcase, the Dart team built Swarm, a slick newsreader. Unfortunately, we only have it in source code right now: http://www.dartlang.org/samples/index.html
As for some "killer" features, I would say there are a few pretty interesting ones:
- optional types are slick, they add annotations and documentations for humans and machines.
- Isolates is a great way to achieve concurrency in a safe manner.
- Libraries (modularity) is sorely needed for the web stack, and Dart has libraries and classes.
- Snapshots will allow for extremely fast start-up
- Bundled libraries (like collections, Stopwatch, etc) will unify code bases and shrink shippable code
- Nice new DOM interface, which makes working with DOM much more enjoyable. It feels like native Dart code.
I hope I've answered your questions. I think the only question that matters is, "Does my language help me build complex, high fidelity, low latency, modular, modern web apps?" The end state of all of this is simply helping more app developers deliver more successful apps to the modern web.