I mentioned in these groups several years ago that trying to port Java to Haiku was putting too much energy into something that wouldn’t yield any great long-term benefits. Java may be good for enterprise back-end progrmming but it has a very poor track record for GUI app development. You can see now how this effort is treading water and will likely just fade away.
It makes far more common sense to have a complete Mono port. Someone ported monolite to Haiku about a year ago which was accepted into the main baseline. That’s a good start.
Writing C++ programs are so error-prone over writing in C# were you can be at least one order of magnitude more productive in your app creation time plus you get a nice development environment with MonoDevelop for free with a full Mono port. You really do have to consider that C++ was the cutting edge when the BeOS came to market. That time is over. .Net apps can access so many powerful libraries in enough interoperable languages to please virtually anyone.
The core Haiku developers should wake up and realize this:
You will get more application developers targeting Haiku if you A) achieve a full Mono port and B) write a P/Invoke layer over the Haiku core C++ library set. Otherwise, I think all you have is a super nitched hacker’s platform which can’t attract enough mainstream developers.
P.S. A 2 gig blank virtual disk isn’t enough space. This project should offer pre-made 10, 20 and 30 gig downloadable virtual drives. Putting the onus on “drive-by programmers” (such as myself) is a real turn-off. Make it easy and inviting to load up the Alpha release and add an extra drive that’s roomy enough to try and port large pre-existing software packages onto.