Primary Source
Sumber Primer
The thing itself: code, transcripts, raw data. Complete and authoritative, but expensive to load.
A primary source is the thing itself: the code, the transcript, the raw data, the contract, the actual log output. It is complete and authoritative, because nobody summarised it and nobody decided what to leave out.
The cost is size. A primary source is usually far bigger than the question being asked of it, so loading the whole thing fills the context with material that will never be relevant. Reading a hundred-page document to answer one question about its third clause is usually the wrong trade.
The practical resolution is to load a small piece of the source and cite it precisely, so a reader can go and check the exact line rather than trusting your paraphrase. This is also why tools that return file contents with a path and line number are better than tools that return a summary: one is verifiable, the other is not.
Related terms
Secondary Source
An account of a primary source, one step removed. Cheap to load, lossy by construction.
Handoff Artifact
A document used as the carry mechanism for a handoff: written by one session, read by another.
Context Engineering
Deciding what goes into the model's context, and what stays out.
Grounding
Tying the answer to something checkable, so it can be verified instead of trusted.
Citation
A pointer from a claim back to the material that supports it.
Want this working inside your team?
It starts with one conversation about your workflow.