Give Each Organizing Tool One Job

If you’ve ever felt your notes (or files, or bookmarks) slowly become unnavigable despite real effort to organize them, the cause is usually not too little structure — it’s redundant structure. Folders, tags, and links are three different organizing tools, and trouble starts the moment two of them encode the same information. The fix that’s kept my own system sane is boring and effective: give each tool exactly one job, and don’t let them overlap.

Three tools, three different questions

Folders, tags, and links aren’t interchangeable ways to do the same thing — they answer different questions:

• Folders answer “what kind of thing is this?” A thing lives in exactly one folder; it’s where the item is physically filed. Folders are good at one place, one answer.
• Tags answer “what is this about, and what state is it in?” They’re flat, multi-valued metadata — a thing can have several — and they’re for filtering and discovery, not primary structure.
• Links answer “what is this related to?” They express relationships between items directly, and they’re the real semantic structure of a well-connected system.

When each tool sticks to its own question, they compose into something powerful. When they blur, you get noise.

Folders say where it lives. Tags say what it’s about. Links say what it relates to. Three questions, three tools — keep them from answering each other’s.

The mess is redundancy, not absence

The specific anti-pattern: encoding the same fact two ways. If a thing already lives in a “networking” folder, slapping a #networking tag on it adds nothing — it’s the same information in a second place, and now you have two systems to keep in sync and twice the chance they drift apart. Duplicated meaning is what makes an organizing system feel heavy and untrustworthy. You stop knowing which signal is authoritative, so you stop trusting any of them.

The discipline is to ask, before adding any tag or folder: does this encode something the other tools aren’t already saying? If the folder already says it, the tag is clutter. If a link already captures the relationship, a tag trying to do the same is redundant.

Links scale where folders don’t

The deeper reason to keep the tools distinct is that they scale differently. Folder hierarchies get worse as they grow — every new item forces a “which branch?” decision, the tree deepens, and things that relate to several topics can only live in one place, so they hide. Links scale the opposite way: an item can be reachable from many others without being copied, and the structure gets richer as it grows instead of more rigid.

So as a system gets bigger, the move is to lean harder on links and lighter on deep folders. A shallow folder scheme for “what kind of thing” plus a web of links for “how things relate” beats an ever-deepening tree every time. The folder tells you the item’s nature; the links let you actually navigate by meaning.

Index notes turn links into navigation

Raw links are powerful but can feel like a pile of threads with no starting point. The thing that turns them into real navigation is a curated index — a note whose whole job is to link out to a set of related notes, grouped and described, giving a topic a front door. Instead of remembering which folder something’s in, you start at the relevant index and follow links. Indexes are how a link-based system stays walkable as it grows: deliberate entry points, maintained by hand, rather than relying on folder paths you have to memorize. They’re also where you notice gaps — a topic with no index, or an index pointing at nothing, tells you where the structure needs attention.

Assign the jobs, then stop overthinking

The payoff of giving each tool one job is that placing things gets easy. Folder choice collapses to “what kind of thing is this?” — usually a one-second answer. Tags become a light dusting of topic and status, not a parallel filing system. Links and indexes carry the meaning. No more agonizing over whether something goes under three different possible folders, because the folder is only answering one small question and the links handle the rest. It’s the same principle behind organizing by what you’ll do with something and building a knowledge system that compounds: clear, non-overlapping roles beat elaborate structure you won’t maintain. If you’ve untangled your own folders-versus-tags-versus-links knot, I’d love to hear how you split the jobs.