For a long time, there are multiple efforts for the concept of "single sourcing": One code base; runs on Web and Desktop without any/much change. I've always strongly rejected the idea of it. Checking out the promising demos of the technology, the pointy-haired Boss simply dreams of the 50% cut in his development cost by choosing the technology. He can be easily convinced that the users don't have to be trained for two different UIs to achieve the same thing right, resulting in cutting cost in documentation, training, bug-fixing etc. Well, fact is far beyond that and the Dilbert in you will understand the practical difficulties.
Web and Desktop are two different beasts. What does it mean for a right click in the browser? Or what does it means for a Back/Refresh button in the Desktop? Both the platforms have their own metaphor. You can't simply throw a Web UI to the Desktop or vice-versa. That's why whenever someone shows me demo of "the brand new framework" for creating both UIs with single source, I would say, "wait for an year for this framework to die". Within an year, there would be an another framework attempting the same.
Orion is different. Its the Web IDE from Eclipse, yet it doesn't try to "leverage" the existing Existing IDE code and present the same UI. Instead its written from the group up. Few months back, when I saw a demo for the StyledText for the web, I was thoroughly impressed. It could handle huge files with ease - which is the bottleneck of other online IDEs. Now that its out to the public as Orion, I'm seeing a lot of +ve reviews and few -ve questions as well. One of the primary question is "How many developers need a Web IDE?". Well, I don't have an exact count. But take away is simple. The market for Web based IDEs is brewing up (eg: Skywriter, Cloud 9) and you would see many more of these in the near future. Just as Eclipse became *the* IDE, Orion is going to be *the* Web IDE. Period.