I suppose this merits Reg coverage
... more or less solely because of the Red Hat patronage. Otherwise it looks like Yet Another JVM Language. That's fine - the nice thing about targeting JVM (or CLR, or LLVM1) is that you can treat these as DSLs when they do something particularly useful in some odd corner of your source base, and use straight-up Java for your main development. It all interoperates and if your development team is any good, maintenance shouldn't be an issue because these languages are highly expressive and a competent developer should be able to pick one up as necessary. (That's assuming the existing code reasonably good, but if you're not holding code reviews and ensuring quality, you get what you deserve.)
But personally I find this one not very interesting. Given world enough and time I'd play around with it, but it ranks pretty low on my list - certainly below Scala and Clojure (which are already pretty well established) in the JVM-languages league, and below Julia and R and some others outside it.
* Compiles to Javascript - OK, but 1) there are other languages that can be cross-compiled to Javascript (including Java), and 2) I'm not keen on the idea anyway. Use Javascript for Javascript, that's my motto. If your script is so damn complex that you need strong typing and class-based OO, then you're using the wrong execution environment.
* HTML syntax - I don't generally do UI work, and when I need HTML, I write HTML, so again this fails to excite me. Even if I did a lot of HTML, Ceylon's support doesn't look too thrilling. What does it do for me that the DOM doesn't? And if I'm generating enough HTML to make it worthwhile, I'm not going to do it with individual method calls for individual DOM nodes; I'll build a higher-level abstraction, thanks.
* Reified generics, type inference, etc - sure, that's nice, but why not go to Clojure and get a proper functional-OO language?
1Or UCSD p-System, right guys? Right? C'mon, where are all the p-System fans?