This translation has not been editorially reviewed yet. The German version is authoritative. Deutsche Fassung
Currency
Why should your website’s HTML pass W3C validation?
HTML is the language in which every website is delivered. The W3C, the body behind the standard, operates a tester for this: the Nu Html Checker. It reads the HTML the way your server actually delivers it, and reports any location that violates the rules of the standard — a non-closed element, a duplicate identifier, an attribute that doesn’t exist at this location. The software test area runs exactly this tester and reports the number of errors together with examples with line indication.
Browsers silently repair — and everyone else
Erroneous HTML is rarely noticed by visitors because browsers correct it when loading: A missing end tag is added, a wrongly nested element is hung. However, this repair is not standardized. Chrome, Firefox and Safari come to different results for the same error, and auxiliary technology like a screen reader or the crawler of a search engine reads something else again. A page that works only thanks to repair works only where you randomly checked it. The validator reveals what the browsers hide: it tells which places had to be repaired at all.
What the test area measures and what it does not claim
The raw HTML of the start page is checked, as it is delivered by the server to the first retrieval – not the version that the browser then assembled. Only errors are counted; Notes and warnings of the validator do not count. The rule “HTML passes W3C validation” carries 15 points in the software status and is only scored if the validator has actually responded. If it could not run, the score remains unchanged – a technical failure on our site is never counted as a defect of your website. The result indicates the number of errors and up to eight examples with row number; It doesn't say the page looks broken because of that, and it doesn't replace accessibility checking.
Valid HTML is the basis on which everything else is based
Search engines, auxiliary technology and translation tools rely on a clear structure. A duplicate identifier makes jump links and form labels go blank, a false nesting tears up lists and tables for the screen reader, an invalid attribute is simply ignored. Many findings from the test areas of accessibility and SEO have their cause in precisely such places. Whoever keeps the HTML valid eliminates sources of error before they appear as their own defects in other check areas – and makes the website regardless of how graciously a particular browser is repairing.
- Have your agency fix the above examples in the source text of the template — There are usually few recurring patterns that disappear on all sides at the same time.
- Check again after each change to the template or after a plugin update; New extensions often bring their own mistakes.
- Pay particular attention to duplicate identifiers (id) and unclosed elements — both hit screen readers and forms directly.
- In your maintenance contract, request valid HTML as an acceptance criterion; The public W3C validator is free for everyone.
