Multiple dungeons in close proximity - campaign setting concept

I'm sort of going to jump on the recent Dark Souls bandwagon here and talk about some stuff I liked about Dark Souls 1.  In addition to all the cool level design that has been discussed on other gaming blogs and forums extensively, one thing I thought was really nice was having a central 'hub' area of the world (Firelink Shrine) where the player can access multiple different areas.  This has got me thinking about how something like this might work in a TTRPG setting or adventure.  Imagine a campaign setting where there are several large dungeons all in close proximity to each other with a central safe area that the players can use as a base.  This allows them the option to basically explore multiple dungeons at the same time, as there would be nothing wrong with them exploring a few rooms of one dungeon before moving on to the next one and coming back to the first one later.  In adventures where a dungeon is the only interesting thing for miles around, the players are just going to try and explore as much of it in one go as they can.

There are lots of different ways you could implement this, e.g. the central hub is a city, and the multiple dungeons lie beneath the city.  I came up with an idea for this concept that I think could be really cool, but would be a lot of work to design and write.  The idea is that the adventure/campaign takes place in a castle that was built by giants.  All the buildings and fortifications are at least 3 times bigger than human-sized ones, meaning that the whole thing would be the size of a small city.  The giants have long since been driven out but other equally-unsavory creatures have moved in to the abandoned buildings.  The enormous bailey and maybe some other parts of the castle are relatively safe, there might even be other civilised humans living there, but the major buildings, underground tunnels/sewers and especially the great keep are all highly dangerous dungeon environments.  Since the buildings are enormous compared to the human PCs, each one would be sizable dungeon that the beasties now inhabiting them would have built their own structures on top of.

There are a few things about this concept that would make it an interesting and unique setting.  The different dungeon areas could be interconnected in a number of ways due to their proximity, adding to the Dark Souls vibe.  Since most of the structures in a castle complex are above ground, there's a lot of room for player creativity in finding different entry points into the dungeons.  The setting would also function to an extent as a megadungeon, but without the players being confined to the dungeon for the whole campaign and having the option of leaving to pursue other adventures and coming back later.  I can just imagine the players getting deep into one of the dungeons and discovering a passage or section that can only be accessed with something they have to go into a completely different dungeon to retrieve.

A huge project like this has a number of major issues that might be easy for an experienced game designer to solve but difficult for an amateur like me, so here's some problems and possible solutions:

Problem: Size and scope of the project.  Something like this will require a huge amount of time dedicated to not only mapping out each area but designing interesting NPCs, monsters, factions and puzzles to fill them with.
Solutions:

• Collaborate with other people to increase the amount of work done over time.
• Do one area at a time and release each one sort of like a little zine.  The issue with this is that the whole concept of the multi-dungeon layout won't come to fruition until after the first few areas are completed.
• Reduce the scope of the project, have only a small number of dungeon areas and have each be fairly small in size.

Problem: Backstory and realism.  There has to be a reason for why the players will want to explore the castle and for why there are hostile entities residing there.  The standard monsters that just live in dungeons because it's their natural habitat (oozes, undead etc.) will get boring very fast if that's all there is in the whole complex.  The setting needs to feel realistic and not as if all the monsters are just placed there for the sake of killing or getting killed by the PCs.

Solution:

• Start the project by designing the history of the castle and surrounding lands.  This is a whole lot of extra work but will lead to a more organic layout of the areas and relationships between NPCs, factions and monsters.  Will probably make another post soon with my basic ideas for this if I go ahead with the project.

Problem: Where do I start?  With a project as big as this I don't really know what to tackle first.  Broad strokes and then fill in the details?  Or start small and work my way out?

Solution:

• Make an action plan before starting anything else.  This will ensure that I always know what I need to work on next.  Here's my basic idea of an action plan for this project:

1. Work out history of the setting.
2. Identify important NPCs and factions.
3. Make a list of dungeon areas.
4. Assign themes and purposes to all areas, keeping in mind the factional interplay.
5. Identify the most important features of each area, including locations and inhabitants.
6. Design a very simple layout of the whole complex and the major points of physical connection where you can get from one to the other.
7. Begin mapping each dungeon, starting with basic layouts and locations of important features and progressively adding more complexity and detail.
8. Before each dungeon is fully stocked, begin adding features that are narratively or functionally connected to features in other dungeons.
9. Finish stocking each dungeon, and make sure there are reasons for the PCs to want to explore every part of each area.  Treasure, knowledge, shortcuts, access points, special items, important NPCs, you get the picture.  Make sure the existence of these things are hinted at in other parts of the complex.
10. Write up any random tables that are needed; a lot of stuff can be left up to random rolls such as monster inhabitants of less important rooms and mundane treasure.
11. Play test.

If I find the time to get started on this massive undertaking I will post my progress on here.  At this point I've only gone as far as surfing through google looking for information on medieval castles, but I don't think that is really going to help me get this project off the ground.  Now is the time to stop procrastinating and start writing!  Wish me luck.