Well there you might have a potential solution. Why not just use project notes for this? Project notes can be accessed from all documents in the Inspector, so once you stick some text in there in one document, you can get at it from another document entirely.
And incidentally there is already a universal place where you can do pretty much what you are describing, the scratchpad. It is a floating panel that can be accessed even outside of Scrivener so it is great for jotting down notes. It has a facility for appending a selection to an existing document, too. Any document you like.
It sounds almost as though you don’t care if the note texts are with the context at all, and having them all in their own place would be more beneficial. If that’s the case, I would recommend using a simple pointer based system. I used to use something similar myself. You just put something unique into the annotation itself—say the date and time (Scrivener makes this easy with Cmd-Shift-Opt-D), and then use that same marker in your “Notes” document (or Project Notes if you prefer). That way, you can get back to the context from the notes document really easily, and you have all of your remarks in one spot easily edited and revised.
Example Chapter Document: Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. [color=red][2008-07-22@11:54]. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
Example Notes Document: 2008-07-22@11:54 — Here is the note text.
I’m not sure if that appeals to you; it doesn’t to me anymore. That is what I was using before Scrivener came along. Now working right in the context is what I prefer, and occasionally producing an exported list of all my notes is good enough for a master list.
I do still mark each annotation with the date and time, however. It makes cross-referencing easy, and also serves as a handy time bomb. I can address things before I forget what I was talking about if I keep working with a certain interval of time. Cross-referencing helps with the issue of resolution chains, as you brought up. Sometimes fixing one thing in chapter 12 fixes eight problems in prior chapters. If I give a certain issue a resolution number, then I can find all eight of them rapidly using the annotation finder with string matching, and delete them.
I’ve always just imported the exported list back into the Binder. 
What you are talking about here would be even more complicated than what I was originally thinking. I am not the developer, so I could be off on my estimation, but I would imagine that an interface allowing remote edits of what essentially amount to specially formatted sections of multiple documents throughout a project would be very complex indeed, and prone to all manner of synchronisation problems.
Again, consider that an annotation or footnote in Scrivener is nothing more involved (technically speaking) than a set of four words set to italic. Appending text to a “clipping document” via the operation of creating an annotation from an existing portion of text would be one thing, as you described in your example, but this example does not address what happens when you turn annotation/footnote marking mode on and then proceed to type in some text from scratch. How is Scrivener to know when you are done editing that remark? When you move out with the cursor, turn the mode off, click to another spot? I can think of plenty of scenarios where I’ve done just that while editing a comment. I’ve even moved to other documents and back again to check on something before finishing. Say I cross-check while typing a note, when I start back up typing again, how does Scrivener know that is the same note? If it just assumes, then it starts to make errors in assumption when a user places to completely separate notes next to each other. The notion of having a clipping document track the text contents of an annotation or footnote area is one of those ideas that seems simple to describe, but from a programming standpoint it is a nightmare.
And that is just one complexity in the most simple method described: Appending to a clipping document. Live database-style dynamic collection with the ability to edit notes as in say an Edit Scrivenings mode for notes would be, as far as I can tell from an outsider’s point of view, practically impossible. I’m sure there is a solution, but it would not be very practical in other words. The main problem would be synchronisation. Since notes are text in a stream, you cannot use byte positioning as the removal of a single character anywhere in the section prior to the note would move it. Byte positioning calibrated every time you click on this hypothetical index page would only work so long as the user was very strict about not using splits. Considering that the operation we are talking about would be exceptionally heavy in just that kind of editing, it would strip a lot of the power from the technique. To put it another way, notes are often used to remind us of how to treat a section of text. If we are going through a list of notes and editing sections of text, those very edits will de-synchronise the position of the notes and Scrivener will lose track of what went where unless it was constantly recalculating and that would grind nearly any computer to a halt. Typing in Leopard is bad enough all by itself.
It’s very messy. Edit Scrivenings works because it is taking whole documents and pasting them end on end in a single window. What you are suggesting is taking only small slices of those documents and pasting them together into one window. There is no magic in a text file that would make this easy, it’s just a long stream of characters (even paragraph breaks are just another character), so re-assembly of a document in slices would require quite a lot of effort. You can’t use string matching instead of position because you might have fifty annotations that all say “SCI: Research” where the context is more important than the words of the note. Try to visualise a text document as a long string of beads. You are wanting to extract 9 beads from a strand of 8,000 where the bracelet designer is constantly changing the beads on the string, and then hoping that Scrivener can get those nine re-arranged beads back into the same spot, even though the new string doesn’t look at all like the old string.
Hopefully that made sense. Again, I’m not privy to what is actually going on, but from what I know of the process, I’m pretty sure my explanation here is accurate. I wouldn’t hope for anything along these lines any time soon, and would instead highly recommend finding a process whereby you can find a way to work within the existing set of tools. There are a lot of existing tools, too. You might come up with something using Scratchpad, Project Notes, References, who knows.