Remember Where, Not What
Senior engineers remember where to look, not every doc. How search tools find things, why a team tag list splits, and what is worth remembering at all.
A senior engineer does not memorize every doc. They know the map, not every street. What you remember is where to look and a few key words.
How search works
| Search | How it finds things | Weak point |
|---|---|---|
| Editor file search (VSCode) | your letters in the same order, with gaps: mysrdme finds my-service/README | you must half-know the name |
| Mail search (Gmail) | whole words, plus filters like from: and subject: | part of a word usually finds nothing |
| Full-text search engine (Elasticsearch) | an index that points each word to its documents. It ranks rare, repeated words higher and allows small typos | it only finds words that someone wrote down |
| Code navigation (LSP, treesitter) | understands the code and jumps to the exact function | works only for code |
The more direct the link, the less you must remember. Simple beats clever here.
The gap in tags
We hoped for one clear tag list for the whole team. But we already name one thing in many ways.
- One doc type, three names:
[Design Doc],[FE Design Doc],[BE Design] - One topic, two titles: “Github Migration” and “GitHub Migration SOP”
A tag list splits the same way, unless everyone memorizes it and someone maintains it. This is the same naming problem as in A Good Name Is a Doc You Never Write.
Is it worth remembering?
- Product: the folder already holds it.
- Words: search already finds them.
- Doc type: Diátaxis sorts docs into four types with two questions. Does the reader want to do something, or to understand it? Are they new to it, or using it at work?
Do you want to remember four names, or two questions? Every tag is one more thing to remember.
Options for your team: a team tag list, doc type only, doc type plus a few topic tags, or no tags with folders and search.
Every tag asks everyone to remember one more thing.
References
Related: revisit the naming post, read who you write for, or go back to the series overview: Docs in the AI Era.