For programming bounties, this would of course need to attract funding from those willing to see a full AMOS reimplementation.
I see 3 stages concerning jAMAL (not necessarily in strict order):
1) Complete the core implementation of AMAL in Java, including full AMAL syntax support, error checking, a debugger/monitor, more efficient use of memory, fully working translation/compilation to Java, and direct support for AMOS BASIC AMAL .abk banks (not to be confused with sprite banks). It should also be able to test if code is forward compatible with the original AMAL (allowing for possible expansion otherwise).
If it is, perhaps it could have the option to launch WinUAE or PUAE directly and test the AMAL scripts on a 'real' Amiga.
2) Improve the graphical engine so it supports multiple Amiga-style screens and screen clipping (currently only one screen is supported), as well as simulated rainbow/copper effects, and native support for AMOS BASIC .abk sprite banks (using one or multiple banks or combined with other sprites). It should also support screen drawing/graphics. This part is already implemented in Jamagic in my "Project Myqu", but will take some work to complete in Java using JGame and/or AWT.
At this stage the jAMAL project would be mostly complete, and a third stage would be as follows:
3) Create a BASIC interpreter in Java, syntax compatible with AMOSPro, and integrate with jAMAL's AMAL interpreter/compiler as well as its graphical engine (and AMAL editor).
Finally, if a native non-Java build is required:
4) Port jAMAL from Java to C/C++, using an appropriate sprite engine, including builds for AROS and classic AmigaOS.
(This could then either port a BASIC interpreter written in Java, or merge with code from Alvyn and/or Mattathias.)
Last edited by Mequa; 06 August 2011 at 04:53.