The following is valid ‘HTML 4.01 Transitional’ according to the W3 validator:
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.or/TR/html4/loose.dtd">
<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<meta name="revisit-after" content="30 days">
<meta name="DC.Title" content="Website title">
<title>Website title</title>
</head><body></body></html>
When transforming this code to HTML5, the meta-tag underwent some changes as documented here. Thus, the following should be valid HTML5:
<!DOCTYPE html>
<html><head>
<meta charset="UTF-8">
<meta name="revisit-after" content="30 days">
<meta name="DC.Title" content="Website title">
<title>Website title</title>
</head><body></body></html>
Except that it doesn’t validate as apparently meta tags are supposed to be registered now.
Problem: The W3 documentation does not list restrictions on meta-tags as a new “feature” of HTML5, but they do not validate like they did previously in HTML 4.01 Transitional.
Update: While the official HTML4 documentation does indeed not restrict the field values of the name attribute, the HTML5 draft mentions the new restriction (unlike the “differences” guide). Some posters suggest to not use meta-tags at all based on SEO arguments, but there have been many public and internal uses of meta-tags for cache control, documentation and storage purposes. Should there not be a way to turn valid HTML4 code into valid HTML5 code without relying on millions of meta-parsers to rewrite themselves automagically?
Question: What should we do in practice?
In practice, just leave the meta tags as they are. Even if the validator complains, it doesn’t make a single bit of difference to anyone using your website.