> Wouldn't it be wonderful to be able to write generic aplets? There are
> many cases where the generic model fits better than the class model.
Undoubtedly. Any reason why that couldn't be done with jgnat?
> And don't worry that AVM may not be able to do JVM work better. There are
> now many examples of software written in Ada that performs better than the
> same software written in C. Ada language features, such as slices for
> arrays, really do make it possible to improve performance over featureless
> C-class languages. (Improve performance in the generated code, that is.)
Those interested in possibly extending the JVM ought to look into
the Latte project, which has produced an open-source JVM/JIT compiler.
The University of Soeul (Korea) is doing this with funding from IBM.
Bsically a research project but they claim big performance improvements
over the Sun JVM. Anyway it's open-source.
Might be interesting to try running jgnat output on this JVM clone.
IMHO it's foolish for the Ada community to reinvent wheels. We are too
small, and our pockets are too shallow, for that. Building and improving
on existing work will be far more fruitful. jgnat, for example, is,
IMHO, just the right kind of idea for Ada to leverage on existing work.
Anyone serious about Ada and Java-ish stuff would do well, in my view,
at least to push the limits of jgnat (open-source, as we know) and
the various open-source Java work, before going off to develop yet
another niche product that can't be sustained for lack of long-term
> S. Ron Oliver, semi-retired professor of Computer Science and Computer
> Engineering. www.csc.calpoly.edu/~sroliver