W3C home > Mailing lists > Public > www-validator@w3.org > June 2017

Re: Closed </li> element

From: Jukka K. Korpela <jukkakk@gmail.com>
Date: Fri, 30 Jun 2017 13:55:02 +0300
Message-ID: <CAGHxYa7h3EzfRd6+raFqKvtiOXBHhzasLYFVkA4cvwFH3+-bQQ@mail.gmail.com>
Cc: "hexel@tut.by" <hexel@tut.by>, W3C WWW Validator <www-validator@w3.org>
2017-06-30 13:35 GMT+03:00 Jens Oliver Meiert <jens@meiert.com>:

> I have found an issue. Validator does not pay attention to unclosed <li>
> > element if it goes after closed </a>.
> It does, it’s just perfectly valid HTML.

More exactly, it is valid (there is really no such thing as being
imperfectly valid) in any version of HTML except those based on XML. The
incantation <!DOCTYPE html> specifies HTML5, and the validator seems to
treat this as HTML5 in HTML serialization, where indeed </li> is always

In XML serialization of HTML5, end tag omission is not allowed. I cannot
find any way to tell the validator to use XML serialization rules when
using direct input. But if you validate a document via URL and the HTTP
headers specify an XML content type, then XML serialization rules will be

Should you wish to do that, you would need to use the xmlns attribute, with
<html xmlns="http://www.w3.org/1999/xhtml">
instead of plain <html>.

Then you would get the error message

*Error*: XHTML element li
allowed as child of XHTML element li
this context. (Suppressing further errors from this subtree.)

It’s somewhat cryptic (as XML validation messages often are), but it
clearly points at the issue.

> (Omitting *all* optional tags can be a very effective way of
> simplifying HTML, but it’s not a wide-spread practice.)

I have no statistics on this, but omitting omissible end tags was common
practice when the Web was young. It was XHTML advocacy that made people
think they gain something by using omissible end tags.

Received on Friday, 30 June 2017 10:55:39 UTC

This archive was generated by hypermail 2.3.1 : Friday, 30 June 2017 10:55:45 UTC