Discussion on "Stop Choosing an Obsidian Template. Engineer a Vault Architecture."

Daisuke Masuda · 2026-10-25T20:20:47.406Z

Most Obsidian setup discussions begin with the wrong question: Which template should I install? That is like choosing a software architecture by comparing starter repositories. A starter can reveal

When an ADR is superseded, how do you flag the runbooks and knowledge notes that still depend on it? Keeping the old decision is useful, but I would want the current instructions to show that they need another review

Yes—that is an important distinction. I would keep the superseded ADR as historical evidence, but change its metadata to something like status: superseded and superseded_by: ADR-0042.

For runbooks and knowledge notes, I use an explicit field such as depends_on: [ADR-0017] or a direct link to the ADR. A review query can then surface every current note that still depends on a superseded decision. I would not automatically mark those notes as wrong—the operational guidance may still be valid—but I would flag them as review_required until someone verifies them against the replacement ADR.

That preserves the decision trail while making the possible blast radius visible. It is also a good example of why queries should be treated as views: the durable truth is the explicit dependency recorded in each note, while the query simply shows which dependencies now require attention.

I bounced between PARA and a zettel-style pile for a year, and the vault got slower to trust than a plain notes folder. The line that landed is that queries are views, not truth. A small folder contract plus a couple of YAML fields made retrieval calmer without adopting a famous template. The Spring Boot latency walkthrough is what makes the hybrid feel like real work.

That’s exactly the failure mode I wanted to address: once the vault’s structure becomes harder to trust than a plain folder of notes, the system is working against you. Treating queries as disposable views—and keeping truth in a small, explicit folder and metadata contract—makes hybrid approaches much less fragile.

I’m especially glad the Spring Boot latency walkthrough helped. I wanted the architecture discussion to survive contact with real engineering work, where one incident can involve logs, hypotheses, source material, decisions, and reusable knowledge without all of them becoming the same kind of note. Thanks for articulating this so clearly.

Search Hashnode

Stop Choosing an Obsidian Template. Engineer a Vault Architecture.

Responses4