[JSR308] Can we agree on our goals?

Michael Ernst mernst at csail.mit.edu
Fri Feb 2 22:11:09 EST 2007


Hani-

> > @LoopMax(10)
> > while (i < j)
> > {
> >  ...
> > }
> >
> > That doesn't seem awkward at all.
> 
> Sure it does. It's a redundant new syntax that doesn't achieve  
> anything, and attacks a problem no developers are currently having  
> (no matter what background they're from or their level of ability,  
> I'd be surprised if anyone found basic looping constructs tricky).  

Several people have given uses for loop annotations on this list.  Examples
include for worst-case execution time analysis and formal methods (proofs,
etc.).  You seem to be dismissing all these people, and saying that
worst-case execution time analysis is of no value, proofs of program
properties are of no value, etc.  If you believe those things, then it
would be more informative to say so explicitly and (especially) to explain
why you hold that position.

I am not here arguing for loop or statement annotations -- they are not
currently part of any JSR 308 proposal, I haven't yet decided whether they
are a good idea, and I am still waiting for a concrete proposal for them.
However, I would like to keep our discussions civil and make sure that each
remark contributes to a productive discussion.  Thanks for your
cooperation.

                    -Mike



More information about the JSR308 mailing list