Jay , I’m one who pushes for document-level settings/properties. I find the metaphor of “documents” extremely useful. Compartmentalization is very important to me. It is, probably, the one thing that I most desperately need Legend for. I’m not excited about the possibility of losing that. Although I would indeed like to draw from across multiple documents, even then I don’t want to draw from ALL of them. And I certainly don’t want my stuff lumped together all the time. I’d rather give up the possibility of sourcing panes from multiple documents than have to deal with a single mega-outline.
This is relevant for organization, for faster navigation, and to aid focus. Legend is specifically useful for me because it lets me handle a bunch of disparate categories and types of information without them cluttering each other. I can’t be distracted by something if I can’t see it. If one huge Outline was the answer to my problems, I would never have come looking for Moo.do. Hell, that’s just a big to-do list.
@LauraH mentions separate sets of tags for each document and I do leverage that aspect of the current Document structure. It is, in fact, what prompted me to start agitating for other things to be per-Document. I don’t want work-tags popping up when I’m trying to auto-complete an item related to personal topics. Same with Move-to destinations when using Alt+M. Documents let me implicitly control the context of my work, so everything can be fast and efficient. Although I will say that a “Tag manager” – requested by many others and comprising a whole range of requested features – could solve that for a “node based” solution by explicitly assigning groups of tags to be “active” under a particular node. But really, how much sense does that make? One big mega-document where users are literally applying different rules to different nodes of it? That defies pretty much every intuitive analogy of digital work. And then you’d just need managers for everything else, to portion it out to nodes. That’s a whole new tier of UI. I know I’m not being very even-handed here but the thought of a mega-document repels me in a visceral way. I promise though, I’ll still listen with an open mind to the arguments.
Sharing/collaboration was one reason I have pushed for per-document settings in the past. It seemed like being able to specify viewing and editing permissions to a whole document would be mandatory to allow collaboration and template-sharing. Assigning this on a document level seemed intuitive and practical. After all, that’s how pretty much everything else out there is shared. By document. There’s several other settings (some you already mentioned) that also make excellent sense to assign per-document, and they’d pair well with the needs of collaboration. In contrast, the idea of sharing on a per-node basis seems, frankly, ridiculous. That’s just too fine of a resolution…and how would that even work into the interface? How would you show a top-level overview of sharing settings across all nodes? What about making changes for several nodes at once? I don’t want a bunch of sharing B.S. in my way while I’m working in a document. That should be something I can assign at a high level and then just take for granted as I work.
Compartmentalization can also have professional implications. For example, I keep all client-specific tasks/information in a single document for that client. I can take that all offline if I stop working with them, provide it directly to the client upon request, or delete it entirely if necessary – without the risk of including something irrelevant. While tools could of course be provided to handle this by-node…why? The current metaphor works. The only exception I feel the need to mention is the Archive command, which has always struck me as something that’d make more sense at the Document level than Item level.
As with the original/instance structure of mirrors, if all my Documents simply become top-level nodes of my account’s single database “under the hood” – but still act like discrete documents when I open the Overview drawer – then I guess I don’t really care. But, also like mirrors, they need to ACT like discrete documents on the surface without belying that true structure. And also-also like mirrors: Let’s beta test the hell out of that change before making it live!