[JSR308] New versions of JSR 308 documents and software

[Neal Gafter]

I still don't see how the proposal disambiguates, in a type annotation such as

    @annotation A.B.T variable;

whether the annotation is annotating the type A, A.B, or A.B.T. At the
risk of belaboring the point, without a specification (including a
grammar) there isn't enough information to evaluate the proposal. I
understand the importance of the informal description, but a precise
specification is also critical.

The proposal fails to mention the disadvantages of the current
proposal: program readers must remember which annotations are type
annotations and which are not, which will be confusing for the readers
of programs. It also fails to mention that the prefix syntax reuses
existing locations for variable declarations, thereby introducing a
syntactic ambiguity. (In fact, the proposal says "the prefix syntax
reuses existing locations for variable declarations without
ambiguity", which is self-contradictory).

I'm working on a code review of the jsr308 prototype. I'll send that separately.

Regards,
Neal

On Jun 7, 2007 1:55 PM, Michael Ernst <mernst at csail.mit.edu> wrote:
>
> We have released, at http://pag.csail.mit.edu/jsr308/#Download , new
> versions of
>  * the JSR 308 design document
>  * the JSR 308 compiler (the prototype implementation)
>  * compiler plug-ins (checkers)
>  * annotation file utilities
>
> There are also descriptions of what is new in the current versions.  For
> example, changes to the design document are listed at
>   http://pag.csail.mit.edu/jsr308/changes.html
>
> As always, both suggestions for further improvements, and contributions of
> improvements (such as patches for code or new text for documents) is
> welcome.  Thanks to all those who have contributed so far.
>
>                     -Mike
>
> _______________________________________________
> JSR308 mailing list
> JSR308 at lists.csail.mit.edu
> https://lists.csail.mit.edu/mailman/listinfo/jsr308
>