The Kava programming language

Hello fellow /prog/ users.

I'm creating a programming language of my own: Kava.

I'd appreciate it very much if you could give me some feedback.

>compiles into bytecode
>platform independent
>executed by a stack-based VM
>based off Java but introduces a lot of changes
>multiple inheritance
>operator overloading
>cleaner syntax
>more light weight
>completely libre
>will support AVR's

My main concern is whether or not should I keep classes as the main construct of the language and the compilation unit. One could say that this forces the OOP paradigm, but I don't agree. I removed interfaces and other OOP-ish constructs inherited from Java, and I feel that keeping the classes in the center will make the language easier to use and easier to implement (and, depending on the person, easier to think in).

Please critique my design and tell me about the problems with my design.

The Kava bytecode is ran by the Kava Virtual Machine (KVM).
I suggest you find a better acronym, instead of overloading another.
>will support AVR's
Are you kidding me? An 8-bit MCU running a VM language? Many embedded devices are slow enough as-is. Unless you're talking about another "AVR"... see the perils of acronym overloading?