Single Comment

> Umm... Here's a stupid question: Cannon to F3, couldn't the 
> rook just capture the cannon, as would happen in orthodox xianqi?

Yet another one who doesn't know where F3 is??? (But at least you got the nature of the question right! ;-) )

But seriously, I don't think this is a viable game. Obviously the mate-in-one problem would have to be solved, but the material imbalance is huge, and even when the rules would be changed to convert the instant mate into an instant Queen-for-Cannon trade, the Xiangqi side should get its *** whipped. So fixing the mate-in-one problem by giving the FIDE side the first move is not really an option.

An interesting way to connect the 9-wide board half to the 8-wide half would be to offset them half by a file (Xiangqi is supposed to be played on the intersections, after all), and add the rule that you should 'keep right' when crossing the River. That would connect the XQ central file (blocked by the Pawn) to the FIDE King file, and the King would be safe. (So you would move a Cannon to d3 to gain a Queen in stead, which would become an obligatory starting move. But you could still choose which Cannon to move there, so it still makes sense to have that move be part of the game.)

By having the square-to-gridpoint connections different for up-rank and down-rank moves, you get the interesting effect that the color-boundedness of Bishops disappears: they can change sqare color by crossing the River, and then crossing back!

connectivity on down-rank moves:
  r n b q k b n r
  p p p p p p p p
  . . . . . . . .
  . . . . . . . .
. . . . . . . . .
P . P . P . P . P
. C . . . . . C .
. . . . . . . . .
R H E A K A E N R

Connectivity on up-rankmoves:
r n b q k b n r
p p p p p p p p
. . . . . . . .
. . . . . . . .
. . . . . . . . .
P . P . P . P . P
. C . . . . . C .
. . . . . . . . .
R H E A K A E N R