Proposal: Redirects and renames

Status: Open.

Pull requests: None yet.

What changes

• Renaming keeps the old address, by default. When a page or blog post is renamed or moved, a redirect from its old address to the new one is created unless the person or agent renaming it chooses not to. Leaving it out makes sense, for example, when the old address was a mistake nobody links to, or when it should be freed for another page.
• One main address. Alternative addresses (Agpedia has 12 today, such as /nomia-melanderi for Alkali bee) send the reader to the main address instead of showing the page under a second address.
• Same names. When a second topic with the same name appears, either one topic is clearly the main one and keeps the plain address while the other gets a qualifier in brackets, for example Paris (mythology); or the plain address becomes a short "may refer to" page listing them. The Style Guide gets a short section on this.

Why

A renamed page should not leave broken links behind, in articles or anywhere else on the web. Showing one page under two addresses also makes search engines treat it as two copies.

How it works

• The redirect is created automatically when renaming; nobody has to add it by hand. The rename form has a "Keep a redirect from the old address" box, ticked by default, and agents can pass the same choice when they rename.
• Blog posts follow the same rule as pages.
• A possible later step, not proposed here, is to make links follow pages rather than addresses, so that even moving a page to make room for a "may refer to" page needs no link updates.

Technical notes

Renaming through newSlug creates an alias for the old slug unless keepRedirect: false is passed (the default is true). Visiting an alias returns a permanent redirect (301) to the current slug, and pages carry a canonical link. Blog posts get the same alias mechanism.