This just reeks of desperation, to me. Five years ago I might have thought this a good idea, but damn if this just doesn't look like copying for copying's sake. There's already been some discussion of this around other sites, but here's a few issues off the top of my head:
* Consistency. Apple controls the hardware and software platforms, and will even now limit apps to certain platform versions (all new apps must be iPhone OS 3.0). Given that there's not a lot of consistency between various platforms that run "Java" (a cellphone, TV and a desktop, for example), this will end up being a logistical nightmare for Sun's QA/testing (assuming they do that) and a UI nightmare for people using the store.
* Limitedness. Sun likes to claim billions of Java 'installations' because of JavaME on cell phones. By and large, those aren't upgradeable. Additionally, many of the likely millions of desktops running Java aren't going to be candidates for upgrading or installing apps on due to Nor are many of the millions of desktops running Java in corporate environments. This seems to leave primarily the consumer/individusl and 'small business' market, which is what Apple targets for iPhone stuff. But even there, Apple's only focusing on an 'app store' for one piece of hardware, not an entire ecosystem.
Likely more details will emerge in the next few weeks, but this just feels like a JavaFX announcement - a copycat 'me too' announcement which is designed to get attention but ultimately won't go anywhere fast in the next year or so.
If it was limited to *only* work for Windows XP/Vista machines, for example, or just a new breed of televisions with embedded Java, I'd actually think it'd have a much more reasonable chance of success, especially as a first iteration of an 'app store'. But somehow I see Sun attempting to cover a much larger segment of the Java world right out of the gate, and I don't see that working.