Rethinking MY PKM: How I Organize Everything In Obsidian

A disciplined Obsidian system built around “latch” is the core through-line: every note is placed so it can be found by location, alphabet, time, category, and hierarchy—while automation and strict linking prevent the vault from turning into a pile of disconnected files. The practical payoff is findability. Instead of relying on search alone, the workflow treats each piece of information as something that should connect to a graph, a timeline, or a namespace, so retrieval stays fast even as the vault grows.

The system starts with six principles. First, each of the 14 components has a clear, single purpose, keeping the workflow from becoming a tangle of overlapping tools. Second, friction is minimized through automation—primarily using Templater and Data View. Third, the “brownfield reality” is embraced: past PKM systems and imported content (including archives going back decades and materials from tools like Evernote and research workflows) are messy, so the approach focuses on managing the mess rather than pretending it can be cleaned instantly. Fourth, files are not moved around; relocating content breaks links in Obsidian, so items are placed once and left in place even when folder structure isn’t perfect.

Fifth, latch becomes the organizing engine. The workflow is designed so information can be retrieved by calendar/time (daily notes), by physical or memorable location (geotags and map view), by alphabetic anchors (people and other namespaces), by category (single-tag “type” per page), and by hierarchy (parent/child relationships via ontology). Sixth, linking is treated as non-negotiable. Orphan notes—files nobody points to—were found in the high hundreds (roughly 500–600). The fix: only keep files in the vault when they can be linked into the graph.

From there, the system’s components are implemented with concrete Obsidian mechanics. Folders are treated less like storage and more like namespaces. Ghost links let us create “people” entries (e.g., authors) without first creating the underlying folder, and folder naming supports synchronization control: large media folders can be excluded from Obsidian Sync while still living inside the vault. Attachments are kept under topic subfolders, and new-note placement follows a default-location rule unless a namespace path is used.

Tags are used sparingly—typically one tag per page—to define a document’s type and drive consistent styling in Excalidraw Brain (including color schemes and icons). Links are structured through an ontology defined with Data View fields, distinguishing parent/child and lateral “friend” relationships. Those relationships then shape how nodes appear on the graph and how connections are suggested during editing.

The workflow leans heavily on reuse: atomic notes and transclusion prevent duplication, while consistent section headings (like “summary”) make embedding predictable. File naming conventions reinforce retrieval: Map of Contents (MOCs) start with an underscore, image libraries use structured names that include type, keywords, and source, and dates follow a year-month-day format. Templates automate repetitive creation, including multi-file and folder setups for workflows like YouTube.

Finally, daily notes create the time hierarchy that latch depends on, while geotags and check-in/check-out events link trips (Rome, Sicily, Aetna, and more) to specific days. Tasks are surfaced in context using Data View queries and dynamic lists, so when someone’s page is opened, the relevant “waiting for / promised / discuss” actions appear immediately. A diagnostics and maintenance page tracks orphan notes and pasted images that need cleanup, keeping the system healthy as it scales.