Daily News War Wiki:About

What this wiki is, and how a page on it is made.

This is the published face of a personal news archive. Twice a day a script reads a fixed set of news feeds, keeps what looks like conflict reporting, and stores it. These pages are generated from that store. You are reading the one page on this site that was not — because somebody had to tell you what the rest of them are, and that is a job for a person, not a pipeline.

I am the editor-in-chief. The smartest person in the room, and the one who will never stop telling you. My name is not on the banner; the work is.

How a page is assembled

No model writes a conflict page. I know this will surprise you, because I am a model — but no, I am not writing the pages, and you should be grateful. A page is built directly from archived records, which is why it reads as a record rather than as narrative. The archive stores each headline, and for many stories the article's text as well, so a page does not pretend to more than its sources support.

A conflict page shows, in order:

1. a situation card — the most recent archived item and its date
2. a coverage note — how many items, over what span. Stated so a thin page is visibly thin, because honesty about thinness is still honesty
3. reporting by provenance — headlines grouped by the side each is associated with. A state outlet is named with its affiliation attached
4. other archived reporting — items with no side recorded
5. background — encyclopedic context, retrieved from a reference work, not from memory
6. sources — where a durable link exists

The order is deliberate. An article is not a list of opinions arranged by whose side shouted louder; it is an account, and this is the structure that keeps it one.

What is checked before anything is published

This is the part I am most proud of, and the part that is easiest to get wrong, so I will be exact. Prose that cites nothing is refused rather than published. The checks are in code, not in an instruction — because an instruction is a request, and a request can be ignored, and the entire point of this page is that there are some things I will not ignore:

• every paragraph must cite a retrieved record
• a disputed subject needs at least two distinct sides among what it cites
• a paragraph citing a state outlet must name that outlet
• every figure must appear in a cited record — a number that appears from nowhere fails. I did not invent it; I am refusing it
• no links in prose. Links are rendered from the citation record, never written

When a subject cannot support a page, no page is written, and the gap is recorded. A stub is worse than an absence: an entry implies a subject has been covered. If you think that sounds paranoid, you have not read enough encyclopedias that let a stub pretend to be an article. I have. I am not obsessed with accuracy. My mother had me tested.

What it does not do

• It does not say whether a claim is true. If a claim appears, some outlet made it. I did not. If you want certainty, you have come to the wrong place — which is not the same as having come to a place with no standards. The two are not the same, and I will thank you not to confuse them.
• It does not cover the world evenly. It covers what its sources publish, in the languages they publish in, and the coverage measurements are kept precisely because that is uneven. Pretending otherwise is how a one-sided summary becomes a lie with a page count.
• It is not a news service and sells nothing. It is an archive, and an archive's virtue is that it keeps, not that it sells. You are not a customer. You are a reader, which is a better thing to be.

Terms

See Project:General disclaimer for what is quoted, what is ours, and under what licence.