Proposal: Automatic Resource Management

Mark Mahieu markmahieu at
Sat Mar 7 08:19:33 PST 2009

Yes, well a key advantage of the kind of library based approach that  
Neal advocates is that it is much easier to improve or evolve after  
the core language change has been made.

But I'm not convinced that the introduction of ARM now would have  
such a reduction in value of a control invocation feature added in a  
later release.  Firstly, the uses of such a feature are far more wide- 
ranging, and secondly, there's the possibility that it could actually  
*replace* the dedicated language feature without requiring any user  
code changes.

If that sounds far-fetched, look at the 'for each' support Neal has  
built into BGGA.  He almost has it at the point that it could  
silently replace the dedicated language support for the for-each loop  
added in Java 5, so although adding BGGA would add complexity to the  
language spec, it could conceivably take some away as well, and give  
us the flexibility of a library-base solution.

The same could conceivably happen with ARM, but it's extremely  
unlikely while the various parties involved are apparently only  
interested in discussing either/or scenarios (yes, I'm guilty of that  



On 7 Mar 2009, at 15:45, Stephen Colebourne wrote:
> My personal take is that I would welcome ARM in Java 7, either
> interface-based or keyword-based. I think the benefits are  
> sufficiently
> great that we shouldn't make developers to wait any longer.
> I am of course fully aware of the reduction in value that BGGA/JCA  
> style
> control invocation then faces as a language change (and hence the
> passion in the debate, notably from Neal).
> However, with closures officially defined as 'out' I have to choose
> between solving real issues now (in the 80% case) or waiting another 3
> years+. I choose pragmatism - solve the 80% now with ARM, and  
> reconsider
> BGGA/JCA later.
> Stephen

More information about the coin-dev mailing list