Things I'm thinking about doing in the near future:
- kill the console Java port
1) It's slow
2) It only works on *nix systems
The .NET port runs *much* faster, and can be used with Mono on
Windows, Linux, and OS X (I've not tried Mono under BSD systems).
The real killer is that Java has no decent console API to work with.
To get key breaking, I have to execute external, non-portable code,
which I can't guarantee will work properly. .NET is much friendlier
for a console developer.
- eliminate the initial_image() embedded in the vm
1) While convenient, it does encourage using out of date images.
2) The vm implementations will become smaller, and builds faster
3) Simpler build process
This would be easy to do, and be a return to the original distribution model.
- split the vm implementations into a separate repository
- split the library into a separate repo
These two would be beneficial in terms of simplifying the overall
codebase (and hopefully becoming more open to additional developers)
- updates to the core image
1) all canvas related words moved out of the retro.forth into canvas file
2) extensions, etc consolidated
-- crc