Method · legacy URL decisions

A redirect needs evidence, not nostalgia.

This is the method used to decide whether a material URL from the former PerlCoders site should return a new page, redirect to a true equivalent, remain missing, or be declared permanently gone.

The acquired domain did not transfer the former company, its customer relationships, software rights, authors or support obligations. That boundary comes before SEO. A useful destination must be independently written, rights-safe and equivalent to the old user intent.

Evidence ladder

What can support an action

Strongest first
LevelEvidenceWhat it can establish
VerifiedDated page capture, current server response, repository test or owned analytics exportA former topic, current behaviour or measurable demand.
SupportedSeveral independent references that agreeA bounded claim, with the source and remaining uncertainty visible.
HypothesisProduct name, URL shape or a single weak referenceA research lead only. It cannot justify invented product details.
RejectedGeneric era assumptions, copied archive prose, unrelated backlinks or a plausible-sounding guessNo public factual claim and no redirect decision.

Five possible actions

  1. Restore as an original 200 page only when the current page serves the same task with new wording and present-day safeguards.
  2. Redirect permanently only when one current URL is a close replacement for the same named resource or intent. A convenient homepage is not equivalent.
  3. Return 404 when no replacement exists and permanent intentional removal cannot be established.
  4. Return 410 for retired executable, account, commerce, download or community surfaces that must not return.
  5. Hold when evidence or rights are unresolved. Holding is a valid outcome; guessing is not.
Current reviewed matrix The repository records 17 material URL rows: twelve precise restore or redirect decisions, three 410 decisions and two 404 decisions. Reviewer and approval timestamps remain empty until a real independent review happens. The routing tests assert known targets and reject an unknown product value.

Rights and identity gate

The current publication may state a product name, a historical URL and a newly written neutral description when the evidence supports them. It does not reproduce old copy, code, downloads, screenshots, logos, customer names or author identities. It never implies that the former operator endorses or continues this publication.

Review checklist

  • Record the exact source URL, query value and known variants before normalising anything.
  • Separate archive evidence from backlink evidence; a link proves a URL was cited, not what the page said.
  • State the current user job and prove that the proposed target performs it.
  • Check rights and identity risk before considering link equity.
  • Test the source response, target response, Location, canonical, robots state and sitemap inclusion.
  • Keep unknown URLs out of catch-all redirects and preserve a correction path.

Sources and reproducibility

The public routing policy follows Google's distinction between permanent redirects and missing pages, while the exact decisions come from the repository manifest and dated evidence attached to each row.