Tim Blackman wrote:
On Mar 29, 2009, at 2:35 AM, Peter Firmstone wrote:
I'd also like to hear some more from Tim Blackman on the issue?
I don't have a strong opinion, but here are some thoughts.
I have been developing (non-Jini/River) code with the Java 5 APIs for
some time now, and have more recently switched to Java 6. There are
definite advantages to switching. Parameterized types are nice, but
not earth shaking. On the other hand, the java.util.concurrent
package is very helpful.
But I also ran into a case, just last week, that we had some code that
we wanted to port to Java ME, which didn't support some Java 5
features we were using. This is not Jini code, though, and I think it
is probably a good argument that most people won't be running Jini on
small devices. (There's a *very* long history behind this issue!)
Still, it would be really too bad to prevent current users of Jini on
JDK 1.4 from adopting River.
I would imagine that Retrotranslator would only add support for
language features, not the currency utilities. Of course, that
probably doesn't matter unless Jini were to be updated to use these
utilities.
- Tim
Yep, it does support Concurrency Utilities, using a backport from Java
6. It also implements additional method parameters in classes in the
Java 5 API, there's a nice table summarising the features here, PS, it
uses ASM ; ) It supports enough of the Java 5 API to be interesting. I
figure that if we use it from the moment we go Java 5, we'll know what
additional Java 5 API features we want that aren't supported and add the
missing functionality to Retrotranslator. Until such a point in time
that it we are certain that we no longer need to support Java 1.4.
http://retrotranslator.sourceforge.net/
