Well I'm happy to consider options ... but there are reasons I absolutely need to keep it as a self contained module, I'm building it for specific purposes where the other features of your module wouldn't necessarily be required. Although, I'm sure some
of them could end up being useful :)
I'm wondering if you're reading too much into my description of it as an "API" - all I mean is that there are a couple of interfaces and some conventions to follow to use the system. This requires
at most a dependency on the project and a reference from the project; which is just the same as you're doing every time you use core Orchard functionality (which I assume you'll be doing a lot!)
In fact the key feature you're after here - drag and drop zone editing - will happen automatically once you enable my module and relevant feature. It just hooks into the relevant parts of the Orchard pipeline. No references or dependencies required; and
you can even use the Recipes system in Orchard 1.1 if you want to have the module automatically download and install along with your bits.
The thing is, the real power of Orchard lies in the ability to compose a website from a large number of individually useful but cunningly interlocking parts. By going against this design choice and trying to build everything into one uber-module you will
actually deprive yourself of the profound flexibility and extensibility that the system naturally provides... so I'd strongly recommend you reconsider your standpoint and look instead at ways to break your own ideas down into separate units that can be used
on their own, but combine together into a much bigger whole. I'm primarily looking to the core modules of Orchard as examples of how this works amazingly well.
Since I'd be distributing this independently anyway, hopefully as open source (although there are specific reasons that might not be able to happen in this case), I'd think we should look at unique ways your platform could leverage this; so there's still
some added value that's unavailable elsewhere. There are literally loads of things you could apply it to and I'm sure something nice could emerge once we've got a better overall understanding of each other's system...
So I hope this makes sense, as in general what you're proposing sounds really interesting and an ideal candidate to really make the most of this feature. :)