UPDATE:
with some help from Qbix (thanks!) and his marvellously documented C++,
I've found where the mouse was getting misdetected, and fixed it.
I'm still testing, there are some remaining calibration issues when the pointer hits
the edge of the screen, but most mousy games should now be playable.
[/b]Dungeon Master:[/b] now runs fine- though sound is a little jerky
(its playable without sound, though nowhere near as much fun)
I suspect EotB should be playable too, though I never liked it to begin with.
I don't think keymapping will fix this.
Sure thing
Remap "keypad-8" to "up arrow"
Remap "keypad-2" to "down arrow"
So, pressing "up" will send a KP_8 to the DOS program. Then you can:
Remap "F1" to [Fn] 1
etc
The problem isn't that "games like this are too complex for the Zaurus"
its that the game UI designers built with a 101-key monstrosity in mind,
and the idea that that some keyboards could be smaller and have
fewer or more keys than the PC-101, was apparently too difficult
for designers back then.
Don't confuse their lack of imagination for a defective Zaurus.
Mouse doesn't work in dosbox, it is 1/3 of the screen away from it's hotspot
Qbix and I (well, Qbix has all the brain power, and I have all the motivation)
are trying to resolve this. About 80% there, for pdaXrom.
In the event the last remaining bits of the problem *are* from pdaXrom or its SDL libs, the mouse may be fine on other roms.
would the first gear head work, maybe in x/qt? It seems to rely on a lot of keyboard commands, which might help or hurt us.
Ew... YUCK
Its PASCAL (the horror!)
but apparently built for FreePascal, which apparently runs on Zaurus.
It uses SDL, no X11 that I can see, so it might just build and run fine on the Z.
You'll need a native build of FreePascal; and whatever SDL dev support it needs.
I can see no reason it wouldn't work, but probably some assembly required.