Our editorial methodology:
useful before impressive.
How the STRUT Field Guide is planned, written, checked, and kept current — including what we refuse to publish.
Map the writing journey
We organize coverage around six jobs: writing with AI, developing notes, structuring long-form work, revising, sustaining a writing practice, and using Strut. A guide is only commissioned when it answers a question one of those jobs actually raises.
State the boundary
We separate durable writing methods from product-specific claims. Sensitive or changing claims need a current primary source. Product descriptions are checked against the shipped software, not the roadmap.
Write for action
Every guide offers a model, a worked example with real before-and-after text, common failure modes, a checklist, and a link to the next useful stage.
Review like an editor
We check structure, factual claims, clarity, accessibility, duplication, and internal links. Overlapping guides get merged rather than multiplied.
Show the trail
Source-dependent pieces preserve provenance. Advice is framed as advice. Every page carries both a published and a last-updated date.
Maintain the library
We update useful pages, deepen thin coverage, and retire material we can no longer support. A guide that cannot be kept accurate is removed rather than left to rot.
What we will and will not publish
Most writing advice on the internet is padded, unsourced, or both. These are the lines we hold, and you can hold us to them.
| We publish | We do not publish |
|---|---|
| Methods we have used on real long-form work | Invented statistics or made-up percentages |
| Worked examples with actual before-and-after text | Research citations we have not read |
| Product claims checked against shipped code | Quotes attributed to people who never said them |
| Guides as long as their argument requires | Filler written to hit a word count |
Why we do not invent numbers
If a claim would need a study to support it, we either cite a primary source or rewrite the claim as reasoning and practitioner experience — which is what most writing advice honestly is. A fabricated statistic makes an article feel authoritative and makes it useless.
Why we do not pad
A guide is as long as its argument requires. We would rather ship a short, complete piece than inflate it, and we would rather deepen an existing guide than launch a thin new one to occupy a keyword.
How AI is used in our own writing
We build an AI writing workspace, so the honest question is whether we use one. We do — the same way we recommend in the AI editing checklist:
- For outlining and pressure-testing structure before drafting
- For finding the unsupported claim in our own draft
- For rewriting passages we already wrote badly
Every published guide is written, edited, fact-checked, and signed off by a person who is accountable for it. Nothing is generated and shipped unread.
How a guide gets made
- Question first. We start from a question a working writer actually asks, not from a keyword with search volume.
- Draft with an example. The worked example is written early, because a method that cannot be demonstrated on real text usually does not survive contact with one.
- Editorial review. Structure, claims, clarity, and whether the piece earns its place beside the guides already in its topic hub.
- Publish with dates. Both dates go on the page so you can judge how current the advice is.
- Revisit. Guides are updated in place when the method or the product changes.
Corrections
If a guide is wrong, tell us at [email protected]. Substantive corrections are made in place and the updated date changes. We do not quietly rewrite history on pieces where the original claim mattered. Product descriptions are re-checked whenever the relevant feature changes.
Who writes this
The Field Guide is written and maintained by the Strut editorial team — the same small group that builds the workspace. That carries an obvious bias worth naming: we think the problems Strut solves are real problems.
We try to earn your trust by making the guides useful to someone who never signs up. Method first, product second, and a clear label whenever a guide is specifically about using Strut. You can read more about the team and the purpose on the about page.
Common questions
Is this content written by AI?
+
No guide is generated and published unread. AI is used for outlining, pressure-testing structure, and rewriting passages we already wrote badly. Every guide is written, edited, fact-checked, and signed off by a person who is accountable for it.
Why are there so few statistics in these guides?
+
Because we do not invent them. A claim that would need a study either cites a primary source or is rewritten as reasoning and practitioner experience, which is what most writing advice honestly is.
How do I report an error in a guide?
+
Email [email protected]. Substantive corrections are made in place and the updated date changes. We do not quietly rewrite history on pieces where the original claim mattered.
Give your rough notes
somewhere to go.
Projects, notes, versioned drafts, imported sources, and careful AI assistance in one private workspace.