That's the major thing that would stop me from using Cappucino over SproutCore... I don't mind learning a new language if there's a compelling need for it, but I just don't see the point here. I'd love to be proved wrong though.
Also, objj is an extension to js, so "learning" it is not some monumental task, kind of like learning the latest additions made by ecmascript 5.
I think there's a tension right now in the web platform community about whether we're in the post-HTML+js world yet. Cappuccino, GWT, jqueryui, etc are all moving in that direction. But there's a tension between libraries that want to treat HTML+js like dumb rendering layer, and libraries that keep HTML+js as the developer-level workspace, and just extend it by adding API.
There are a lot of things that work really well in HTML+js (links, urls, scrolling, etc) that get screwed up as you add more post-HTML widgetry. So the developers of post-HTML tools eventually end up rewriting everything HTML did in their new toolkit. The toolkits get better and better, but I have yet to see one that didn't break something that HTML did well.
So it becomes an either/or proposition. Either I jump fully into Cappuccino, accept that it will break things about HTML and I will not be able to fix them, or I stay in the land of HTML, and I try to implement the cool stuff Cappuccino does (windowing, OO, advanced widgetry, etc) with HTML+js and whatever JS libraries I can find, but generally sick with The HTML Way wherever possible.
The thing is, I know what HTML sucks at. I have no way to know what Cappuccino breaks that HTML does well until I am waist-deep in a Cappuccino app. I think that makes it hard for me to jump.
And so my preference (although it's just a preference) is for small tools that add one thing to HTML+js, but they do it really well. So if I want to make a mobile app for iPhones, I want to start with a JS library that lets me make touch-slideable panes, that's incredibly fast and smooth, and that's it. I don't need it to create a whole complex windowing model, or re-implement HTML buttons in javascript, or anything like that.
But I feel like the tide is slowly turning... the post-HTML frameworks are getting better. At some point one will be good enough that I really can leave HTML behind. Maybe I'll try out Cappucino for my next weekend project and see how it goes.
Cappuccino the framework is open source.