There's also an option to have a shared base layer (showing the basic layout of the land and such, I suppose), and each participant supplies a transparent png (or other image format with transparency - could even be postscript I reckon) with his/her area filled in. The images can then be merged into one map every now and then as collab WIP's and the different parts can be discussed either from the merged image or from the single transparency overlayed on the base image.

I think I'd enjoy doing something like this, though I'd probably end up doing more supporting development work (as with the Context Free script I made for Ravells) - I do hope I'd count for a team member even if I hadn't actually drawn a pixel on the final map.

Also, does it really matter how much each person has participated? I reckon if one person made the whole thing, they'd be rather miffed at the ones that didn't help, and so would want to, say, make note of that on their entry?