Enter Your Reply

Large boards are certainly difficult to work on successfully; with all that
scope, you have much more room for error. And you also run into a problem
of scaling vs. playability. Any sorts of simplistic extrapolations to
large size will run into a host of problems, many of which translate to
tedium. A certain creativity is called for, a walk off the beaten path.
That walk may often end tangled in brambles or floundering in a sinkhole,
but sometimes it will lead to places you only thought you'd see in your
dreams. I've seen some of David Paulowich's ideas. I think he'll come
out with a game that meets his high standards and is in keeping with his
design philosophy. I'll wish him luck, but I doubt he'll need it. I
will, looking seriously at superlarge variants, games in the 20x20 to
30x30 range, just above my posted games range of 8x8 to 19x19. Got some
practice, and think there are some guides to successful [2D] supergames. 
Moving multiple pieces per turn should speed the game up.
Don't get carried away with pieces or piece types. Too many of either
makes the game unplayable. Strict scaling to a 600 square board would give
each side 150 pieces, which is probably ridiculous. Around 50 pieces is
probably a good number as a general rule; this seems manageable.
Balance the pieces to the size of the game. Using standard FIDE pieces and
piece ratios is probably a bad idea. 'Eight of Everything' chess would
fit nicely on a 24x24 board:
PPPPPPPPPPPPPPPPPPPPPPPP
RNBBNRRNBBNRRNBBNRRNBBNR
xxxxQQQQKKKKKKKKQQQQxxxx
All FIDE rules are in effect except:
1 castling, as the Ks and Rs are not aligned for it;
2 victory, which has new  conditions, primarily by capturing all your
opponents kings before you lose all yours;
3 movement, because you must move a different piece for each king you have
remaining on the board each turn, or you lose.
It's even got 64 pieces, fitting nicely with our general principle. But I
don't think it would play very well. It violates too many other principles
to be a realy good game. Enough for now, more later.