
A web page can be reachable at several addresses even though its contents appear exactly the same. A publication might resolve both a route ending in a slash and a file address ending in .html. A query parameter could change nothing about the main document. Old extension-based routes can remain in use long after the publisher reorganizes the site. For a person reading one version, none of this may be noticeable. For a team maintaining navigation, search discovery and site reports, the ambiguity can lead to inconsistent references. Defining one preferred address for the document makes the route inventory easier to understand and gives the site a consistent place to send visitors.
Identify variants that really describe the same document
Do not begin by treating every address with a shared prefix as a duplicate. Consider the difference between '/journal/performance/' and '/journal/performance/?view=compact'. If compact changes only the presentation while preserving the same text and purpose, both addresses may be alternate representations of the same article. A category page containing the full collection and a filtered page containing a distinct subset may serve different purposes. The correct treatment depends on what changes. Compare titles, headings, visible content, meaningful interactions and any reason a visitor would deliberately choose the variant. Classify equivalent variants separately from routes that represent a new or narrower document.
Create a route convention before changing behavior. Decide how informative pages are addressed, whether directory-style pages end with a slash, how article and service routes are composed, and which scheme and hostname the site prefers. For an ordinary static site, the visible address can end in a slash while its internal build contains an index file. The file name does not have to become another advertised page URL. Pick a convention that the server and release process can preserve reliably. It is better to keep an existing widely used convention than to introduce needless movement in pursuit of a visually pleasing but operationally inconsistent format.
Use a permanent redirect when a route has moved
Suppose an article was previously served at '/visual-studies/lines.html' and has been incorporated into '/visual-studies/'. If the new page covers the same material and remains the most suitable destination, a permanent redirect from the old path is appropriate. Configure the old address to direct a request to the final preferred URL, rather than to a second outdated address that then redirects again. Internal links should reference the preferred URL immediately. An old path may still receive requests from bookmarked pages or references outside the site, so preserving the redirect is useful even after no current navigation contains that path.
A redirect changes navigation behavior for the requesting person as well as for automated consumers. After following it, the browser typically shows the destination address. That is desirable for an old route, but it may not suit an alternate representation that has a useful function. For instance, a parameter controlling the order of a genuinely equivalent list may need to remain available so that a visitor can preserve that preference. The site can still identify the preferred document address through canonical markup when the relationship is appropriate. Distinguish an obsolete address from an intentional alternate before deciding that every variant should disappear.
Related reading: A sitemap that reflects the site you actually publish.
Understand what canonical markup communicates
The canonical link element in the page head identifies the preferred URL associated with the document. It is a signal, not a mechanism that forcibly redirects a visitor or creates the destination. A page at an equivalent query variant might include a canonical link identifying its base route. The base route ordinarily identifies itself. The server must still return useful content at that preferred address. Pointing a canonical link to an unavailable page, to a redirect or to a document addressing a different question sends contradictory information. Do not use the element as a shortcut for resolving an unrelated publishing or routing problem.
Consider an educational site with a broadly useful article about generative lines and another one about preserving a procedural art seed. They share vocabulary and some mathematical background, but the second has a distinct purpose: reproducing a specific output later. Identifying the first as canonical for the second would misrepresent the content relationship. If the two pieces are actually redundant, an editorial merger may make sense. In that case the surviving article should incorporate necessary material and the superseded equivalent can redirect. A canonical configuration should not be used to conceal two thin or repetitive pages whose purpose the publisher has never differentiated.

Align the other documents that advertise the route
A publishing system commonly generates several representations of the same route inventory. The category page lists the article, the breadcrumb references a parent, a related reading section links to it, and the XML sitemap lists a preferred address. These producers should draw the URL from the same publication record. If one uses a physical file path, another appends an undocumented slash and a third constructs an address from a title, they can diverge after an update. Store the chosen public address separately from the output location. For a generated article, use that public address in the page head, listings and relationship links instead of independently constructing variants in every template.
Keep the sitemap restricted to the actual preferred pages that the publisher intends to make discoverable. Do not list obsolete paths on the assumption that redirects will eventually lead search systems to the new pages. Similarly, a navigation link to a redirect adds an unnecessary request and makes the source page harder to audit. There may be a few deliberate internal references to a legacy address when it is itself the subject of a technical example, but such examples are not ordinary navigation. Ensure that address-like text presented as an example is not accidentally generated as a link to an obsolete page.
Check HTTP behavior, not only the displayed page
A browser that follows redirects automatically can hide important details. Request the old path in a way that lets you inspect the initial status and location. Then request the destination and confirm that it returns the expected document rather than another redirect or a missing page. Record the preferred URL found in the returned page head. Perform a similar check for alternate combinations of host, scheme, extension and trailing slash if they are relevant to the deployment. The aim is to determine which alternatives actually resolve. A theoretically possible form is not necessarily a route that the site currently serves.
A useful small test table has columns for requested address, expected status class, expected destination, final status and canonical URL. Include a current page, an old equivalent page, a useful parameter variant and a deliberately nonexistent path. A request for a nonexistent article should not silently return the home page with a successful status just to avoid showing an error. If a generated document has a canonical link, check that the URL is absolute or otherwise resolves exactly to the intended address. Then match the result against the sitemap and inspect a few actual internal links, particularly from shared navigation and article summaries.
Keep canonical choices stable as content changes
The address of a useful page can remain stable while its title, design or contents improve. A new heading need not trigger a new slug, especially if the page continues to answer the same question. When a route really needs to change, update the route inventory and the affected redirect mapping as part of that release. Preserve the old mapping instead of creating a chain that depends on intermediate names remaining configured. Review parameters and routes introduced by new interface features. If they expose alternate documents, decide explicitly whether the alternatives remain accessible, should redirect or represent distinct pages. A consistent URL model makes the next editorial or technical change easier to reason about.
Explore the implementation area: Technical SEO.


