Editorial Policy

This page sets out what we publish, what we will not claim, and how the engine behind the site is documented when it changes.

What we will not claim

We will not describe any generated name as best, perfect or recommended. The engine measures spelling. It cannot hear a name, cannot know what it means to you, and cannot know what a community already uses. Ranking by structure is defensible and calling a result best is not.

We will not publish research results we have not collected. The human review study page carries its protocol and states that it has no data yet, and it will keep saying that until there is data.

How research pages are produced

The regression checks and the limits report are generated by calling the live blending engine during the site build. No figure on either page is typed in by a person, so those pages cannot drift out of step with the tool, and a scoring change becomes visible on the next deploy.

The methodology page carries the engine version and is updated whenever the scoring changes, including the weighting choices we made and why.

Corrections

When we get something wrong we fix the page and update its date. Engine behaviour errors take priority over everything else, because the research pages read from the engine and a fault there propagates across the site.

Advertising and independence

The site is supported by advertising. Advertisers have no input into what is published, how the engine scores, or what any page says about its limits. There are no affiliate links and nothing is sold here.

Ad slots appear on guide, research and article pages only. The three tools carry no advertising in their interfaces, because an ad inside a tool that is arguing for careful judgement would undercut the point.