Maruno rotation idea seems interesting, but I thinks that the big problem are that one week can be too much/little to do certain things, may generates problems with continuity, and is too serial. The most effective (in my vision) are some groups with:
In parallel instead of per week.
Each group have a leader and maybe coordinators, with the main development team and only the main leader of main development (who needs to be very active) needs edit the main project. The key is to make separate and parallel things, like sprites that, with a group leader that manage people to don't do the same thing (rework is a great enemy), works very well in parallel, even with dozens of people. With maps and script may work with copy/paste the file/script changes. The main leader only needs to combine everything and done one thing or other, like events, and even may do with models done for other people. Maybe the leader launch some builds (maybe open source) every 1-3 months to the group have a model to follow. Of corse, some people can be in more than one grop at same time.
The project will remain as long that exist people with interest and a minimun skill to pass the torch forward, keeping an active project holder.
Just as I said, the main problem is the lack of interested people with certain skill level. If you divide the work to one people/group to update the tilesets/charsets, one people to update the PBS, one people to do the maps, among others, this can be very more quicker.
Continuity would be sorted out by people discussing it in the project's thread, as well as the occasional tidy-up volunteer (I said I might volunteer to do that). The allotted time per person could be different, or vary according to what each person asks for (to a limit) - I suggested a week as a starting point, since no one knows what the optimal way of working this kind of idea would be - it hasn't been done before.
I honestly doubt a project like this which requires teams and team members will be able to gather them all, let alone retain them. You said this yourself. I don't think there's any advantage to making teams out of people rather than let people make contributions if/when they want. Committed people will work regardless of whether they're called "part of the team" or not.
I've already said that other people can do things in parallel with the current week's volunteer. I gave examples of town maps, Pokédexes and so forth, and that could quite easily apply to maps too. It'd be up to the volunteers to add those contributions into the game (if they want to). I
did notice that you apparently assumed a complete replacement of all the graphics would be
required, which is a rather silly thing to assume.
Besides, the point isn't necessarily about speed (not that I think pass-it-around would be appreciably slower than other methods anyway). It's about it being a community project. "Serial" or not, it lets everyone have an equal go at it. Surely that's the point of calling it "communal"? Don't dismiss it just because it's different, because it
is different and may well require a different approach to every other fangame.
I disagree, even the biggest wikis have their teams and leaders.
This isn't a wiki, this is a game. Wikis automatically allow everyone to contribute regardless of the leadership structure. If Wikipedia only let nominated staff edit articles, and asked the rest of the world to discuss and propose changes (which could simply be ignored),
then it would be more like what you're proposing. It would also be rubbish, obviously, and definitely couldn't be called a community project.
If there is a team and/or leader (preferably just 1 person, if any), then their job should solely be to make sure things don't get out of hand (this
doesn't include preventing the game from turning into something they don't personally like). The less intervention on their part, the better.