About these docs
The Atlas documentation follows Diátaxis: four modes, strictly separated, each with its own voice. If a page feels like it belongs in two modes, it is doing its job badly — please flag it.
The four modes
Section titled “The four modes”Tutorials teach Atlas by doing. They assume nothing, take you by the hand through a worked example, and end with a working artifact. Read these first if you are new to Atlas.
How-to guides solve a specific problem. They assume you already know what Atlas is for and skip the orientation. They are indexed by the UI surface you are working in — Investigate, Cases, Studio, Graph, Compare, Governance.
Reference is the flat, factual side of the site. API endpoints, gate catalog, error codes, ontology, config keys. Reference pages do not explain motivation; they document what is there.
Explanation answers why. The questions that start with why a design looks the way it does, why a number is what it is, why Atlas makes the call it makes.
Audiences
Section titled “Audiences”Atlas speaks to three readers:
- Technical architect — evaluating whether Atlas fits a conversion program. Start with What Atlas does and Evidence and confidence.
- ABAP developer or consultant — working a specific task. Jump straight to the how-to guides and keep the reference open in a second tab.
- Platform integrator — wiring Atlas into CI, ATC, or Datasphere. Start with the API reference and the configuration reference.
Business stakeholders are served by naburis.cloud, not by these docs.
Giving feedback
Section titled “Giving feedback”Every page has a 👍/👎 control at the bottom. Use it. The ratings land in the DJED admin dashboard; we read them, we act on them, and a page with a string of 👎 ratings is how we find out it is wrong.
For anything longer than a thumb press, the edit on GitHub link at the bottom of every page opens a PR against the source file directly.
Reading these docs with an LLM
Section titled “Reading these docs with an LLM”Every page exposes a Copy page action in its header that copies
the page’s raw Markdown to your clipboard, ready to paste into the
model of your choice. The same content is served at
/docs/raw/<slug>.md for scripted pipelines.
Atlas docs are written so this works well — the prose is self-contained per page, code samples are copy-pasteable, and every claim links to the evidence behind it.