Question

I encountered this page https://www.google.com/accounts/ServiceLogin, a Google service login page that (beyond just omitting a doctype), contains 6 instances of </img>

For example,

  <img src="https://www.google.com/accounts/google_transparent.gif"
           alt="Google">
  </img>

Why would they ever do that? What benefit/functionality/grandfathering do they possibly achieve?

Anything I’ve ever read about HTML and XHTML has made it pretty unequovical:

In HTML 4.01 and prior, <img> tags are never to be closed ( <img src="img.gif" alt="text" >).

In XHTML, <img> tags are to be closed using self-closing syntax ( <img src='img.gif' alt="text" />)

In HTML5, (my understanding is that) either syntax (open or self-closed) is acceptable, but still never </img>.

Answer 1

I’ve found the only (proposed) way this code is ever actually compliant, though it does not apply in Google’s case (since they lack a DOCTYPE).

XHTML 2, which was proposed and then scrapped, implements a </img> tag as a way to replace the alt attribute.

So, instead of this in XHTML 1.0/1.1:

<img src="monkeys.gif" alt="Monkeys throwing feces" />

You’d have this

<img src="monkeys.gif">Monkeys throwing feces</img>

Where ‘Monkeys throwing feces’ only displays if monkeys.gif fails to load.

This would make <img> behave as other content embedding tags, like <object>.

In the spec’s words,

The img element is a holder for
embedding attributes such as src.
Since these attributes may be applied
to any element, the img element is not
strictly necessary, but is included to
ease the transition to XHTML2. Like
the object element, this element’s
content is only presented if the
referenced resource is unavailable.