[JSR308] Can we agree on our goals?

[Trevor Harmon]

On Feb 1, 2007, at 10:24 PM, Ted Neward wrote:

>> 1) Once and for all, either come up with a way to support each
>> construction or declare that it cannot (or should not) be done.
>>
>> versus
>>
>> 2) Face a continuing stream of JSR proposals, each providing a
>> solution to one lacking in the previous annotation-related JSR.
>>
> Doug, I hate to say it, but I think this is a naïve distinction;  
> for every X
> that we say cannot be done, somebody will propose a new JSR that  
> shows how
> it can be done, why it should be done, and why we were all idiots  
> for not
> doing it before now.

That's probably true. I also think it would be dangerous to try to  
solve all of these issues in one shot and hope we get it right the  
first time.

So what exactly is wrong with (2)? I'm not seeing the problem in  
having a stream of JSRs. In fact, it seems logical to expand the  
annotation mechanism incrementally, step-by-step, learning as we go.  
A natural progression might be:

Java 5 --> annotations
Java 6 --> pluggable annotations and common annotations
Java 7 --> annotations on types (and loops?)
Java 8 --> annotations on any arbitrary statement

Trevor